[FFmpeg-devel] [PATCH] lavfi/vf_colorconstancy: add weighted_greyedge
Hi,
This patch was part of GSoC Color Constancy filter. It introduces an
improved color constancy filter(weighted_greyedge) based on the already
pushed greyedge. I'm sending it again after updating it against latest
version of master.
Regards,
Mina Sami
>From 33b92a766ff575c9765c15097cdda19a2e9e9e60 Mon Sep 17 00:00:00 2001
From: Mina
Date: Sun, 30 Sep 2018 21:32:17 +0200
Subject: [PATCH] lavfi/vf_colorconstancy: add weighted_greyedge
Signed-off-by: Mina
---
doc/filters.texi| 39
libavfilter/Makefile| 1 +
libavfilter/allfilters.c| 1 +
libavfilter/vf_colorconstancy.c | 322
4 files changed, 288 insertions(+), 75 deletions(-)
diff --git a/doc/filters.texi b/doc/filters.texi
index 6d3833317e..f00ee4781c 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -17526,6 +17526,45 @@ separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
@end example
@end itemize
+@section weighted_greyedge
+A color constancy variation filter which estimates scene illumination via weighted grey edge
+algorithm and corrects the scene colors accordingly.
+
+See: @url{http://refbase.cvc.uab.es/files/GGW2012.pdf}
+
+The filter accepts the following options:
+
+@table @option
+@item minknorm
+The Minkowski parameter to be used for calculating the Minkowski distance. Must
+be chosen in the range [1,20] and default value is 1.
+
+@item sigma
+The standard deviation of Gaussian blur to be applied on the scene. Must be
+chosen in the range [0.2,1024.0] and default value = 1.
+
+@item wmap_k
+The power applied to the weight map to emphasis heigher weights. Must be chosen
+in the range [1,20] and default value = 2.
+
+@item iters
+The max number of iterations for performing the algorithm. Must be chosen in the
+range [1,20] and default value = 5.
+
+@item angerr
+The angular error threshold in degrees for stoping the algorithm. Must be chosen
+in the range [0,360] and default value = 0.001.
+@end table
+
+@subsection Examples
+@itemize
+
+@item
+@example
+weighted_greyedge=minknorm=5:sigma=2:wmap_k=10:iters=10:angerr=0.0005
+@end example
+@end itemize
+
@section xbr
Apply the xBR high-quality magnification filter which is designed for pixel
art. It follows a set of edge-detection rules, see
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 62cc2f561f..f9626d4e16 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -400,6 +400,7 @@ OBJS-$(CONFIG_VSTACK_FILTER) += vf_stack.o framesync.o
OBJS-$(CONFIG_W3FDIF_FILTER) += vf_w3fdif.o
OBJS-$(CONFIG_WAVEFORM_FILTER) += vf_waveform.o
OBJS-$(CONFIG_WEAVE_FILTER) += vf_weave.o
+OBJS-$(CONFIG_WEIGHTED_GREYEDGE_FILTER) += vf_colorconstancy.o
OBJS-$(CONFIG_XBR_FILTER)+= vf_xbr.o
OBJS-$(CONFIG_YADIF_FILTER) += vf_yadif.o
OBJS-$(CONFIG_ZMQ_FILTER)+= f_zmq.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index 5e72803b13..095da84151 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -381,6 +381,7 @@ extern AVFilter ff_vf_vstack;
extern AVFilter ff_vf_w3fdif;
extern AVFilter ff_vf_waveform;
extern AVFilter ff_vf_weave;
+extern AVFilter ff_vf_weighted_greyedge;
extern AVFilter ff_vf_xbr;
extern AVFilter ff_vf_yadif;
extern AVFilter ff_vf_zmq;
diff --git a/libavfilter/vf_colorconstancy.c b/libavfilter/vf_colorconstancy.c
index e3bb39e51b..3b84a637c7 100644
--- a/libavfilter/vf_colorconstancy.c
+++ b/libavfilter/vf_colorconstancy.c
@@ -26,6 +26,14 @@
*
* @cite
* J. van de Weijer, Th. Gevers, A. Gijsenij "Edge-Based Color Constancy".
+ *
+ * @cite
+ * J. van de Weijer, Th. Gevers, and J. Geusebroek,
+ * âEdge and corner detection by photometric quasi-invariantsâ.
+ *
+ * @cite
+ * A. Gijsenij, Th. Gevers, J. van de Weijer,
+ * "Improving Color Constancy by Photometric Edge Weighting".
*/
#include "libavutil/imgutils.h"
@@ -39,7 +47,8 @@
#include
-#define GREY_EDGE "greyedge"
+#define GREY_EDGE "greyedge"
+#define WEIGHTED_GREY_EDGE "weighted_greyedge"
#define SQRT3 1.73205080757
@@ -79,6 +88,10 @@ typedef struct ColorConstancyContext {
int minknorm; /**< @minknorm = 0 : getMax instead */
double sigma;
+int wmap_k;
+int iters;
+double angerr_thresh;
+
int nb_threads;
int planeheight[4];
int planewidth[4];
@@ -477,53 +490,73 @@ static int filter_slice_grey_edge(AVFilterContext* ctx, void* arg, int jobnr, in
}
/**
- * Main control function for grey edge algorithm.
+ * Slice function for weighted grey edge algorithm that is called iteratively to
+ * calculate and apply weight scheme.
*
* @param ctx the filter context.
- * @param in frame to perfrom grey edge on.
+ * @param arg data to be passed between threads.
+ * @param jobnr current job nubmer.
+ * @para
Re: [FFmpeg-devel] Mentoring project: music test source
On 09/30/2018 07:41 PM, Nicolas George wrote: Hi. For the next rounds of sponsored internships, I would like to propose the following project, that I would mentor: A music-like audio lavfi source for testing purposes. That means a deterministic pseudo-random stream of notes with varied frequencies, with a structure that looks like music Do you mean by "look like music" to just follow music theory rules or actually sound like real music. Just curious. and would trigger the same pathways in filters and codecs. It would be based on asrc_sine, which already has a fast bit-exact sinusoidal signal generator. On top of that, the filter would change the frequency randomly to make notes according to a tempo, preferably by following harmony rules, and would shape the note with an envelope and harmonics. As qualification tasks, I can propose: - make a "musicalscale" filter source that outputs the audible spectrum of notes in sequence; - make a filter source that sounds like a percussion instrument hit repeatedly at a regular interval. What do you think of it, in terms of usefulness and difficulty? Regards, ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [PATCH] lavfi/vf_colorconstancy: add weighted_greyedge
On 10/01/2018 12:24 PM, Moritz Barsnick wrote: +The standard deviation of Gaussian blur to be applied on the scene. Must be ^ to Done. +The power applied to the weight map to emphasis heigher weights. Must be chosen ^ emphasize British spelling but done. +The angular error threshold in degrees for stoping the algorithm. Must be chosen ^ stopping Done. + * @cite + * J. van de Weijer, Th. Gevers, and J. Geusebroek, + * ???Edge and corner detection by photometric quasi-invariants???. ^^^ You should fix the quotation marks here. Done. - * @return 0 in case of success, a negative value corresponding to an - * AVERROR code in case of failure. + * @return 0. */ -static int filter_grey_edge(AVFilterContext *ctx, AVFrame *in) +static int filter_slice_weighted_grey_edge(AVFilterContext* ctx, void* arg, int jobnr, int nb_jobs) If it only ever returns 0, you can probably make it void - unless it will be expanded later to require a return code. AVFilterInternal::execute expects such type of the function. Moritz ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
[FFmpeg-devel] [PATCH v2] lavfi/vf_colorconstancy: add weighted_greyedge
Hi,
This patch was part of GSoC Color Constancy filter. It introduces an
improved color constancy filter(weighted_greyedge) based on the already
pushed greyedge. I'm sending it again after updating it against latest
version of master.
V2 changes:
- fixed some spelling mistakes in filters.texi
- fixed some text format error in filters.texi
Regards,
Mina Sami
>From 158159a2137d104763875718f1c217c472b12fa5 Mon Sep 17 00:00:00 2001
From: Mina
Date: Mon, 1 Oct 2018 18:06:41 +0200
Subject: [PATCH] lavfi/vf_colorconstancy: add weighted_greyedge
Signed-off-by: Mina
---
doc/filters.texi| 39
libavfilter/Makefile| 1 +
libavfilter/allfilters.c| 1 +
libavfilter/vf_colorconstancy.c | 322
4 files changed, 288 insertions(+), 75 deletions(-)
diff --git a/doc/filters.texi b/doc/filters.texi
index 6d3833317e..d2b0ef024b 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -17526,6 +17526,45 @@ separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
@end example
@end itemize
+@section weighted_greyedge
+A color constancy variation filter which estimates scene illumination via weighted grey edge
+algorithm and corrects the scene colors accordingly.
+
+See: @url{http://refbase.cvc.uab.es/files/GGW2012.pdf}
+
+The filter accepts the following options:
+
+@table @option
+@item minknorm
+The Minkowski parameter to be used for calculating the Minkowski distance. Must
+be chosen in the range [1,20] and default value is 1.
+
+@item sigma
+The standard deviation of Gaussian blur to be applied to the scene. Must be
+chosen in the range [0.2,1024.0] and default value = 1.
+
+@item wmap_k
+The power applied to the weight map to emphasize heigher weights. Must be chosen
+in the range [1,20] and default value = 2.
+
+@item iters
+The max number of iterations for performing the algorithm. Must be chosen in the
+range [1,20] and default value = 5.
+
+@item angerr
+The angular error threshold in degrees for stopping the algorithm. Must be chosen
+in the range [0,360] and default value = 0.001.
+@end table
+
+@subsection Examples
+@itemize
+
+@item
+@example
+weighted_greyedge=minknorm=5:sigma=2:wmap_k=10:iters=10:angerr=0.0005
+@end example
+@end itemize
+
@section xbr
Apply the xBR high-quality magnification filter which is designed for pixel
art. It follows a set of edge-detection rules, see
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 62cc2f561f..f9626d4e16 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -400,6 +400,7 @@ OBJS-$(CONFIG_VSTACK_FILTER) += vf_stack.o framesync.o
OBJS-$(CONFIG_W3FDIF_FILTER) += vf_w3fdif.o
OBJS-$(CONFIG_WAVEFORM_FILTER) += vf_waveform.o
OBJS-$(CONFIG_WEAVE_FILTER) += vf_weave.o
+OBJS-$(CONFIG_WEIGHTED_GREYEDGE_FILTER) += vf_colorconstancy.o
OBJS-$(CONFIG_XBR_FILTER)+= vf_xbr.o
OBJS-$(CONFIG_YADIF_FILTER) += vf_yadif.o
OBJS-$(CONFIG_ZMQ_FILTER)+= f_zmq.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index 5e72803b13..095da84151 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -381,6 +381,7 @@ extern AVFilter ff_vf_vstack;
extern AVFilter ff_vf_w3fdif;
extern AVFilter ff_vf_waveform;
extern AVFilter ff_vf_weave;
+extern AVFilter ff_vf_weighted_greyedge;
extern AVFilter ff_vf_xbr;
extern AVFilter ff_vf_yadif;
extern AVFilter ff_vf_zmq;
diff --git a/libavfilter/vf_colorconstancy.c b/libavfilter/vf_colorconstancy.c
index e3bb39e51b..28b466f89a 100644
--- a/libavfilter/vf_colorconstancy.c
+++ b/libavfilter/vf_colorconstancy.c
@@ -26,6 +26,14 @@
*
* @cite
* J. van de Weijer, Th. Gevers, A. Gijsenij "Edge-Based Color Constancy".
+ *
+ * @cite
+ * J. van de Weijer, Th. Gevers, and J. Geusebroek,
+ * "Edge and corner detection by photometric quasi-invariants".
+ *
+ * @cite
+ * A. Gijsenij, Th. Gevers, J. van de Weijer,
+ * "Improving Color Constancy by Photometric Edge Weighting".
*/
#include "libavutil/imgutils.h"
@@ -39,7 +47,8 @@
#include
-#define GREY_EDGE "greyedge"
+#define GREY_EDGE "greyedge"
+#define WEIGHTED_GREY_EDGE "weighted_greyedge"
#define SQRT3 1.73205080757
@@ -79,6 +88,10 @@ typedef struct ColorConstancyContext {
int minknorm; /**< @minknorm = 0 : getMax instead */
double sigma;
+int wmap_k;
+int iters;
+double angerr_thresh;
+
int nb_threads;
int planeheight[4];
int planewidth[4];
@@ -477,53 +490,73 @@ static int filter_slice_grey_edge(AVFilterContext* ctx, void* arg, int jobnr, in
}
/**
- * Main control function for grey edge algorithm.
+ * Slice function for weighted grey edge algorithm that is called iteratively to
+ * calculate and apply weight scheme.
*
* @param ctx the filter context.
- * @param in frame to perfrom
Re: [FFmpeg-devel] Mentoring project: music test source
On 10/04/2018 11:40 AM, Nicolas George wrote: Hi. No need to Cc people who are on the list. Actually replied only to you initially, sorry about that. Mina (2018-09-30): Do you mean by "look like music" to just follow music theory rules or actually sound like real music. Just curious. In my mind, the purpose is to have a signal that can exercise filters and codecs, trigger as many code paths as possible, without requiring an external sample. Therefore, it must "look like music" from the point of view of analysis done by filters and codecs. And since some of the have quite subtle psychoacoustic models tweaked for music, following music theory rules is probably a requirement. Sounding like real music seems much more difficult, and possibly not that useful. But I think that if the code is clean, if somebody comes up with a better algorithm to sound more catchy, it can be added in the same filter as an option. And, to answer Carl Eugen's question: No, I do not have a student lined for this project. Regards, I think I can start on that, I was already searching for my next contribution. Pedro's reply seems pretty useful. ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [PATCH v2] lavfi/vf_colorconstancy: add weighted_greyedge
On 10/01/2018 06:09 PM, Mina wrote: Hi, This patch was part of GSoC Color Constancy filter. It introduces an improved color constancy filter(weighted_greyedge) based on the already pushed greyedge. I'm sending it again after updating it against latest version of master. V2 changes: - fixed some spelling mistakes in filters.texi - fixed some text format error in filters.texi Regards, Mina Sami A gentle ping. ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [PATCH v2] lavfi/vf_colorconstancy: add weighted_greyedge
On 11/28/18 11:24 PM, Moritz Barsnick wrote: On Mon, Oct 01, 2018 at 18:09:46 +0200, Mina wrote: Nit: Thanks for your feedback and sorry for late response. +be chosen in the range [1,20] and default value is 1. [...] Not sure what is here. +chosen in the range [0.2,1024.0] and default value = 1. A bit inconsistent here, with "is" vs. "=". Done, changed to "=". +The max number of iterations for performing the algorithm. Must be chosen in the ^ maximum Done, changed to "maximum". I can't comment on the rest. Cheers, Moritz ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
[FFmpeg-devel] [PATCH v3] lavfi/vf_colorconstancy: add weighted_greyedge
Hi,
This patch was part of GSoC Color Constancy filter. It introduces an
improved color constancy filter(weighted_greyedge) based on the already
pushed greyedge. I'm sending it again after updating it against latest
version of master.
V3 changes:
- fixed inconsistency in filters.texi
Regards,
Mina Sami
>From f4891ca38470893d29864821cd1988bb026cf160 Mon Sep 17 00:00:00 2001
From: Mina
Date: Tue, 4 Dec 2018 00:22:23 +0200
Subject: [PATCH] lavfi/vf_colorconstancy: add weighted_greyedge
Signed-off-by: Mina
---
doc/filters.texi| 39
libavfilter/Makefile| 1 +
libavfilter/allfilters.c| 1 +
libavfilter/vf_colorconstancy.c | 322
4 files changed, 288 insertions(+), 75 deletions(-)
diff --git a/doc/filters.texi b/doc/filters.texi
index 41fbbc5329..1cad919ba7 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -18038,6 +18038,45 @@ separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
@end example
@end itemize
+@section weighted_greyedge
+A color constancy variation filter which estimates scene illumination via weighted grey edge
+algorithm and corrects the scene colors accordingly.
+
+See: @url{http://refbase.cvc.uab.es/files/GGW2012.pdf}
+
+The filter accepts the following options:
+
+@table @option
+@item minknorm
+The Minkowski parameter to be used for calculating the Minkowski distance. Must
+be chosen in the range [1,20] and default value = 1.
+
+@item sigma
+The standard deviation of Gaussian blur to be applied to the scene. Must be
+chosen in the range [0.2,1024.0] and default value = 1.
+
+@item wmap_k
+The power applied to the weight map to emphasize heigher weights. Must be chosen
+in the range [1,20] and default value = 2.
+
+@item iters
+The maximum number of iterations for performing the algorithm. Must be chosen in the
+range [1,20] and default value = 5.
+
+@item angerr
+The angular error threshold in degrees for stopping the algorithm. Must be chosen
+in the range [0,360] and default value = 0.001.
+@end table
+
+@subsection Examples
+@itemize
+
+@item
+@example
+weighted_greyedge=minknorm=5:sigma=2:wmap_k=10:iters=10:angerr=0.0005
+@end example
+@end itemize
+
@section xbr
Apply the xBR high-quality magnification filter which is designed for pixel
art. It follows a set of edge-detection rules, see
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 6e2658186d..d77ca697b2 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -414,6 +414,7 @@ OBJS-$(CONFIG_VSTACK_FILTER) += vf_stack.o framesync.o
OBJS-$(CONFIG_W3FDIF_FILTER) += vf_w3fdif.o
OBJS-$(CONFIG_WAVEFORM_FILTER) += vf_waveform.o
OBJS-$(CONFIG_WEAVE_FILTER) += vf_weave.o
+OBJS-$(CONFIG_WEIGHTED_GREYEDGE_FILTER) += vf_colorconstancy.o
OBJS-$(CONFIG_XBR_FILTER)+= vf_xbr.o
OBJS-$(CONFIG_XSTACK_FILTER) += vf_stack.o framesync.o
OBJS-$(CONFIG_YADIF_FILTER) += vf_yadif.o yadif_common.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index a600069500..7dbfccb393 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -392,6 +392,7 @@ extern AVFilter ff_vf_vstack;
extern AVFilter ff_vf_w3fdif;
extern AVFilter ff_vf_waveform;
extern AVFilter ff_vf_weave;
+extern AVFilter ff_vf_weighted_greyedge;
extern AVFilter ff_vf_xbr;
extern AVFilter ff_vf_xstack;
extern AVFilter ff_vf_yadif;
diff --git a/libavfilter/vf_colorconstancy.c b/libavfilter/vf_colorconstancy.c
index e3bb39e51b..3b84a637c7 100644
--- a/libavfilter/vf_colorconstancy.c
+++ b/libavfilter/vf_colorconstancy.c
@@ -26,6 +26,14 @@
*
* @cite
* J. van de Weijer, Th. Gevers, A. Gijsenij "Edge-Based Color Constancy".
+ *
+ * @cite
+ * J. van de Weijer, Th. Gevers, and J. Geusebroek,
+ * âEdge and corner detection by photometric quasi-invariantsâ.
+ *
+ * @cite
+ * A. Gijsenij, Th. Gevers, J. van de Weijer,
+ * "Improving Color Constancy by Photometric Edge Weighting".
*/
#include "libavutil/imgutils.h"
@@ -39,7 +47,8 @@
#include
-#define GREY_EDGE "greyedge"
+#define GREY_EDGE "greyedge"
+#define WEIGHTED_GREY_EDGE "weighted_greyedge"
#define SQRT3 1.73205080757
@@ -79,6 +88,10 @@ typedef struct ColorConstancyContext {
int minknorm; /**< @minknorm = 0 : getMax instead */
double sigma;
+int wmap_k;
+int iters;
+double angerr_thresh;
+
int nb_threads;
int planeheight[4];
int planewidth[4];
@@ -477,53 +490,73 @@ static int filter_slice_grey_edge(AVFilterContext* ctx, void* arg, int jobnr, in
}
/**
- * Main control function for grey edge algorithm.
+ * Slice function for weighted grey edge algorithm that is called iteratively to
+ * calculate and apply weight scheme.
*
* @param ctx the filter context.
- * @param in frame to perfrom grey edge on.
+ * @param arg
Re: [FFmpeg-devel] [PATCH 1/2] libavfilter/vf_colorconstancy.c : Cleanup code for new filter
On 6/18/20 3:52 PM, Yatendra Singh wrote: Signed-off-by: Yatendra Singh --- libavfilter/vf_colorconstancy.c | 47 ++--- 1 file changed, 14 insertions(+), 33 deletions(-) LGTM ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-devel] [PATCH 2/2] libavfilter/vf_colorconstancy.c : Adding weighted greyedge
On 6/18/20 3:52 PM, Yatendra Singh wrote:
Signed-off-by: Yatendra Singh
---
doc/filters.texi| 36 ++
libavfilter/Makefile| 1 +
libavfilter/allfilters.c| 1 +
libavfilter/vf_colorconstancy.c | 215
4 files changed, 253 insertions(+)
diff --git a/doc/filters.texi b/doc/filters.texi
index 85a511b205..2946b5b9e6 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -20250,6 +20250,42 @@
separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
@end example
@end itemize
+@section weighted_greyedge
+Apply the color constancy filter which estimates illumination and updates the
+image colors accordingly.
+
+It accepts the following options:
+
+@table @option
+@item minknorm
+The Minkowski parameter to be used for calculating the Minkowski distance. Must
+be chosen in the range [0,20] and default value is 1. Set to 0 for getting
+max value instead of calculating Minkowski distance.
+
+@item sigma
+The standard deviation of Gaussian blur to be applied on the scene. Must be
+chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} *
break_off_sigma(3) )
+can't be equal to 0 if @var{difford} is greater than 0.
+
+@item min_err
+The error angle between the estimated illumination and white light at which
the algorithm stops even
+if it hasn't reached the required number of iterations @code{max_iters}. Must
be chosen in the range
+[0.02,PI] radians with default of 0.1.
+
+@item max_iters
+The number of iterations at which the algorithm stops even if it hasn't
reached the required
+error angle between the estimated illumination and white light @code{min_err}.
Must be chosen in the
+range [1,100] with a default value of 10.
+
+@item kappa
+The power which is applied to the spectral weights to change the impact of
weights on illuminant
+estimation @code{kappa}. Must be chosen in the range [1,25] with a default
value of 10.
+@end table
+
+@example
+ffmpeg -i mondrian.tif -vf
"weighted_greyedge=minknorm=0:sigma=1:max_iters=50:min_err=0.02:kappa=10"
mondrian_out.tif
+@end example
+
@section xbr
Apply the xBR high-quality magnification filter which is designed for pixel
art. It follows a set of edge-detection rules, see
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 994a4172a3..6973452f8d 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -453,6 +453,7 @@ OBJS-$(CONFIG_VSTACK_FILTER) += vf_stack.o
framesync.o
OBJS-$(CONFIG_W3FDIF_FILTER) += vf_w3fdif.o
OBJS-$(CONFIG_WAVEFORM_FILTER) += vf_waveform.o
OBJS-$(CONFIG_WEAVE_FILTER) += vf_weave.o
+OBJS-$(CONFIG_WEIGHTED_GREYEDGE_FILTER) += vf_colorconstancy.o
OBJS-$(CONFIG_XBR_FILTER)+= vf_xbr.o
OBJS-$(CONFIG_XFADE_FILTER) += vf_xfade.o
OBJS-$(CONFIG_XFADE_OPENCL_FILTER) += vf_xfade_opencl.o opencl.o
opencl/xfade.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index f2a44b0090..ad2e07f9c5 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -432,6 +432,7 @@ extern AVFilter ff_vf_vstack;
extern AVFilter ff_vf_w3fdif;
extern AVFilter ff_vf_waveform;
extern AVFilter ff_vf_weave;
+extern AVFilter ff_vf_weighted_greyedge;
extern AVFilter ff_vf_xbr;
extern AVFilter ff_vf_xfade;
extern AVFilter ff_vf_xfade_opencl;
diff --git a/libavfilter/vf_colorconstancy.c b/libavfilter/vf_colorconstancy.c
index d36400bd35..e2e32b7ca3 100644
--- a/libavfilter/vf_colorconstancy.c
+++ b/libavfilter/vf_colorconstancy.c
@@ -1,5 +1,6 @@
/*
* Copyright (c) 2018 Mina Sami
+ * Copyright (c) 2020 Yatendra Singh
*
* This file is part of FFmpeg.
*
@@ -26,6 +27,14 @@
*
* @cite
* J. van de Weijer, Th. Gevers, A. Gijsenij "Edge-Based Color Constancy".
+ *
+ * @cite
+ * J. van de Weijer, Th. Gevers, and J. Geusebroek,
+ * “Edge and corner detection by photometric quasi-invariants”.
+ *
+ * @cite
+ * A. Gijsenij, Th. Gevers, J. van de Weijer,
+ * "Improving Color Constancy by Photometric Edge Weighting".
*/
#include "libavutil/imgutils.h"
@@ -40,8 +49,10 @@
#include
#define GREY_EDGE "greyedge"
+#define WEIGHTED_GREY_EDGE "weighted_greyedge"
#define SQRT3 1.73205080757
+#define NORMAL_WHITE 1/SQRT3
#define NUM_PLANES3
#define MAX_DIFF_ORD 2
@@ -77,12 +88,16 @@ typedef struct ColorConstancyContext {
int difford;
int minknorm; /**< @minknorm = 0 : getMax instead */
+int kappa;
double sigma;
int nb_threads;
int planeheight[4];
int planewidth[4];
+double min_err;
+int max_iters;
+
int filtersize;
double *gauss[MAX_DIFF_ORD+1];
@@ -608,6 +623,162 @@ static void chromatic_adaptation(AVFilterContext *ctx, AVFrame *in, AVFrame *out
ctx->internal->execute(ctx, diagonal_tr
Re: [FFmpeg-devel] [PATCH v3] lavfi/vf_colorconstancy: add weighted_greyedge
Hi, I have been very busy last year but recently things are settling down and I would be glad to resume contributing. FYI, a contest was held earlier this year that resulted in some great papers/variations, so I can either resend an updated patch or refactor the filter to accommodate for those possible new variations. Cheers, Mina Sami On 9/22/19 3:13 PM, Paul B Mahol wrote: Is this gonna be applied at all? On 12/3/18, Mina wrote: Hi, This patch was part of GSoC Color Constancy filter. It introduces an improved color constancy filter(weighted_greyedge) based on the already pushed greyedge. I'm sending it again after updating it against latest version of master. V3 changes: - fixed inconsistency in filters.texi Regards, Mina Sami ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe". ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-devel] [PATCH] avfilter/vf_colorconstancy: fix overreads in gauss array
On 10/12/19 11:09 AM, Paul B Mahol wrote:
Fixes #8250
Signed-off-by: Paul B Mahol
---
libavfilter/vf_colorconstancy.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/libavfilter/vf_colorconstancy.c b/libavfilter/vf_colorconstancy.c
index e3bb39e51b..cc081e957f 100644
--- a/libavfilter/vf_colorconstancy.c
+++ b/libavfilter/vf_colorconstancy.c
@@ -280,7 +280,7 @@ static int slice_get_derivative(AVFilterContext* ctx, void*
arg, int jobnr, int
dst[INDX2D(r, c, width)] = 0;
for (g = 0; g < filtersize; ++g) {
dst[INDX2D(r, c, width)] += GAUSS(src, r,
c + GINDX(filtersize, g),
- in_linesize, height,
width, gauss[GINDX(filtersize, g)]);
+ in_linesize, height,
width, gauss[g]);
}
}
}
@@ -295,7 +295,7 @@ static int slice_get_derivative(AVFilterContext* ctx, void*
arg, int jobnr, int
dst[INDX2D(r, c, width)] = 0;
for (g = 0; g < filtersize; ++g) {
dst[INDX2D(r, c, width)] += GAUSS(src, r +
GINDX(filtersize, g), c,
- width, height,
width, gauss[GINDX(filtersize, g)]);
+ width, height,
width, gauss[g]);
}
}
}
LGTM
Thanks,
Mina
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
[FFmpeg-devel] [PATCH][GSoC] srcnn - an image super resolution filter using CNN
Hi,
This patch is introduced as a qualification task required by Super
Resolution project for GSoC. It passes patchcheck and make fate and
doesn't introduce new warnings. It's implemented by the help of the
mentor: Pedro Arthur. It's been tested over 7 images of which 6 got
expected results while 1 didn't.
Used command for testing:
ffmpeg -i input_image -vf "scale=2*iw:2*ih,format=yuv420p,srcnn"
output_image.
Attached is the patch and here is a link <https://imgur.com/a/pTZTm> for
3 images with their results; one of which is the one with incorrect result.
Please check(the problem maybe in the way I am copying the image) and
share some feedback or guidance about what the problem may be and what
should I do next regarding the GSoC proposal as I haven't submitted one
yet. Is there any template/requirements for proposal? Also note that the
algorithm runs convolution without multi-threading on cpu as simple
while loops so it's slow for large images. Performance be improved later
on.
Best regards,
Mina Sami.
>From 2ac25e4456a3ce9a16f688cce590da2521ff4703 Mon Sep 17 00:00:00 2001
From: MinaBombo
Date: Thu, 22 Mar 2018 16:53:22 +0200
Subject: [PATCH] Added srcnn filter
---
Changelog | 1 +
libavfilter/Makefile| 1 +
libavfilter/allfilters.c| 1 +
libavfilter/srcnn_weights.h | 290
libavfilter/vf_srcnn.c | 261 +++
5 files changed, 554 insertions(+)
create mode 100644 libavfilter/srcnn_weights.h
create mode 100644 libavfilter/vf_srcnn.c
diff --git a/Changelog b/Changelog
index c1b9df4..e153f12 100644
--- a/Changelog
+++ b/Changelog
@@ -46,6 +46,7 @@ version :
They can be found at http://git.videolan.org/?p=ffmpeg/nv-codec-headers.git
- native SBC encoder and decoder
- drmeter audio filter
+- Added image SRCNN(Super Resolution Convolutional Neural Network) filter. No training yet.
version 3.4:
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index fc16512..b37a956 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -323,6 +323,7 @@ OBJS-$(CONFIG_SMARTBLUR_FILTER) += vf_smartblur.o
OBJS-$(CONFIG_SOBEL_FILTER) += vf_convolution.o
OBJS-$(CONFIG_SPLIT_FILTER) += split.o
OBJS-$(CONFIG_SPP_FILTER)+= vf_spp.o
+OBJS-$(CONFIG_SRCNN_FILTER) += vf_srcnn.o
OBJS-$(CONFIG_SSIM_FILTER) += vf_ssim.o framesync.o
OBJS-$(CONFIG_STEREO3D_FILTER) += vf_stereo3d.o
OBJS-$(CONFIG_STREAMSELECT_FILTER) += f_streamselect.o framesync.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index cc423af..42f409c 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -332,6 +332,7 @@ static void register_all(void)
REGISTER_FILTER(SOBEL, sobel, vf);
REGISTER_FILTER(SPLIT, split, vf);
REGISTER_FILTER(SPP,spp,vf);
+REGISTER_FILTER(SRCNN, srcnn, vf);
REGISTER_FILTER(SSIM, ssim, vf);
REGISTER_FILTER(STEREO3D, stereo3d, vf);
REGISTER_FILTER(STREAMSELECT, streamselect, vf);
diff --git a/libavfilter/srcnn_weights.h b/libavfilter/srcnn_weights.h
new file mode 100644
index 000..6497b1e
--- /dev/null
+++ b/libavfilter/srcnn_weights.h
@@ -0,0 +1,290 @@
+/*
+ * Copyright (c) 2018 Mina Sami
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file
+ * Super resolution filter pretrained weights
+ *
+ * @see https://arxiv.org/abs/1501.00092
+ */
+
+#ifndef AVFILTER_SRCNN_WEIGHTS_H
+#define AVFILTER_SRCNN_WEIGHTS_H
+
+const int ff_cnn_num_filters[] = {
+1, 64, 32, 1
+};
+
+const int ff_cnn_size_filters[] = {
+1, 9, 1, 5
+};
+
+static const float weights1_64x81[] = {
+-0.088663,0.05541,0.037197,-0.11961,-0.12342,0.29963,-0.091182,-0.00013614,-0.049024,0.038421,-0.077268,0.027273,0.45762,0.023582,-0.20364,-0.077384,0.072884,0.066799,-0.073151,0.26653,0.078364,-0.28985,-0.50691,0.044665,-0.062469,-0.024425,-0.029073,-0.16885,0.20686,-0.25367,-0.385,-0.12602,0.72579,0.095884,-0.044873,-0.011252,-0
[FFmpeg-devel] [Patch][GSoC] srcnn - a super resolution filter using CNN
Hi,
This patch is introduced as a qualification task required by Super
Resolution project for GSoC. It passes patchcheck and make fate, doesn't
introduce new warnings and gives expected results for all tested images.
It's implemented by the help of the mentor: Pedro Arthur.
Best regards,
Mina Sami.
>From bf8ac38cee36886f6c317681a16aecfa5dee0752 Mon Sep 17 00:00:00 2001
From: MinaBombo
Date: Sat, 24 Mar 2018 21:15:08 +0200
Subject: [PATCH] Added srcnn filter
---
Changelog | 1 +
libavfilter/Makefile| 1 +
libavfilter/allfilters.c| 1 +
libavfilter/srcnn_weights.h | 290
libavfilter/vf_srcnn.c | 264
5 files changed, 557 insertions(+)
create mode 100644 libavfilter/srcnn_weights.h
create mode 100644 libavfilter/vf_srcnn.c
diff --git a/Changelog b/Changelog
index c1b9df4..e153f12 100644
--- a/Changelog
+++ b/Changelog
@@ -46,6 +46,7 @@ version :
They can be found at http://git.videolan.org/?p=ffmpeg/nv-codec-headers.git
- native SBC encoder and decoder
- drmeter audio filter
+- Added image SRCNN(Super Resolution Convolutional Neural Network) filter. No training yet.
version 3.4:
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index fc16512..b37a956 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -323,6 +323,7 @@ OBJS-$(CONFIG_SMARTBLUR_FILTER) += vf_smartblur.o
OBJS-$(CONFIG_SOBEL_FILTER) += vf_convolution.o
OBJS-$(CONFIG_SPLIT_FILTER) += split.o
OBJS-$(CONFIG_SPP_FILTER)+= vf_spp.o
+OBJS-$(CONFIG_SRCNN_FILTER) += vf_srcnn.o
OBJS-$(CONFIG_SSIM_FILTER) += vf_ssim.o framesync.o
OBJS-$(CONFIG_STEREO3D_FILTER) += vf_stereo3d.o
OBJS-$(CONFIG_STREAMSELECT_FILTER) += f_streamselect.o framesync.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index cc423af..42f409c 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -332,6 +332,7 @@ static void register_all(void)
REGISTER_FILTER(SOBEL, sobel, vf);
REGISTER_FILTER(SPLIT, split, vf);
REGISTER_FILTER(SPP,spp,vf);
+REGISTER_FILTER(SRCNN, srcnn, vf);
REGISTER_FILTER(SSIM, ssim, vf);
REGISTER_FILTER(STEREO3D, stereo3d, vf);
REGISTER_FILTER(STREAMSELECT, streamselect, vf);
diff --git a/libavfilter/srcnn_weights.h b/libavfilter/srcnn_weights.h
new file mode 100644
index 000..d6ac79c
--- /dev/null
+++ b/libavfilter/srcnn_weights.h
@@ -0,0 +1,290 @@
+/*
+ * Copyright (c) 2018 Mina Sami
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file
+ * Super resolution filter pretrained weights
+ *
+ * @see https://arxiv.org/abs/1501.00092
+ */
+
+#ifndef AVFILTER_SRCNN_WEIGHTS_H
+#define AVFILTER_SRCNN_WEIGHTS_H
+
+const int ff_cnn_num_filters[] = {
+1, 64, 32, 1
+};
+
+const int ff_cnn_size_filters[] = {
+1, 9, 1, 5
+};
+
+static const float weights1_64x81[] = {
+-0.088663,0.05541,0.037197,-0.11961,-0.12342,0.29963,-0.091182,-0.00013614,-0.049024,0.038421,-0.077268,0.027273,0.45762,0.023582,-0.20364,-0.077384,0.072884,0.066799,-0.073151,0.26653,0.078364,-0.28985,-0.50691,0.044665,-0.062469,-0.024425,-0.029073,-0.16885,0.20686,-0.25367,-0.385,-0.12602,0.72579,0.095884,-0.044873,-0.011252,-0.029559,0.22117,-0.38453,-0.05351,0.10956,0.37257,-0.28968,-0.10884,0.19629,0.097468,-0.0056471,-0.13036,0.62543,0.19104,-0.011692,-0.27068,-0.030754,0.029876,0.016809,-0.18318,-0.16016,0.35983,-0.1889,-0.22343,-0.2,0.32211,-0.18514,-0.06647,0.18006,-0.03566,-0.25352,-0.1837,0.28395,0.0030237,0.24073,-0.14355,0.047419,-0.032212,-0.088013,0.29592,0.041087,-0.10354,-0.11681,-0.010267,0.023704,
+-0.043954,0.095339,0.12234,-0.052115,0.068812,0.057649,0.086867,-0.12001,0.023536,-0.022271,-0.063779,0.010644,-0.27824,-0.2372,-0.21418,-0.09665,0.10459,0.065994,-0.031174,-0.034654,0.022144,-0.24322,-0.089236,-0.027248,-0.093149,-0.03741,0.078197,0.12521,0.1119,0.34982,0.24304,0.21351,0.37121,0.12193,-0.15492,0.01884,-0.083511,-0.3055,-0.11462,-0.1
[FFmpeg-devel] [GSoC] proposal idea: Color Constancy using Machine Learning
Hi, I'm proposing a Color Constancy filter as a project idea for GSoC and was asking for any remarks about that idea. Color constancy is the ability of the human visual system that ensures the perception of the color of objects to remain relatively constant under varying illumination conditions. A color constancy filter would basically be applied on images with various degrees of damage to restore original colors as close as possible. This <https://imgur.com/a/hj2ox> is a simple example of a low-statistics filter output I've implemented as an academic project along with the original damaged image. The filter is an implementation of a mathematical equation presented with some parameters that's used to introduce variations to the filter. Use cases: despite being an end-result filter that helps retrieving damaged images original colors it's very helpful as a pre-processing utility for many computer vision algorithms. A simple example would be a classifier that identifies an object utilizing its color. Implementation: this is still a research filed where last paper I've read was out earlier this year and still more and more papers are being published introducing various learning-methods to improve accuracy including unsupervised and supervised methods and even new learning-based algorithms to improve low-statistics light-computational ones. Variations of this filter and their accuracies will also prove to be a very appropriate project for GSoC since there is many factors affecting the choice of filter such as real-time need, filed of application, available dataset, ..etc. So, it's more like a group of filters. Impact: introduction of machine learning to this filed is quiet recent so enhancing FFmpeg with this technology and possibly adding to it will prove entrepreneurship. Photography, Computer Vision and many other fields that performs fine image processing would make good use of this filter. Best regards, Mina Sami. ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [Patch][GSoC] srcnn - a super resolution filter using CNN
Hi, I haven't received any feedback regarding this qualification task whether it's accepted or not. So, with less than 10 days left for GSoC, I thought I was ask for any reviews about my task so that I can fix it if anything needs fixing. Thanks in advance. Best regards, Mina Sami. On 03/25/2018 10:27 AM, Mina wrote: Hi, This patch is introduced as a qualification task required by Super Resolution project for GSoC. It passes patchcheck and make fate, doesn't introduce new warnings and gives expected results for all tested images. It's implemented by the help of the mentor: Pedro Arthur. Best regards, Mina Sami. ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
[FFmpeg-devel] [PATCH][GSOC] avfilter: added colorconstancy
Hi,
This patch introduces Grey-Edge algorithm as part of the Color
Constancy Filter project in GSOC.
Best regards,
Mina Sami
>From 75ff45f28690cf208058315e4a3ec48e7981b837 Mon Sep 17 00:00:00 2001
From: Mina
Date: Thu, 12 Jul 2018 20:19:55 +0200
Subject: [PATCH] avfilter: added colorconstancy
Signed-off-by: Mina
---
Changelog | 1 +
MAINTAINERS | 1 +
libavfilter/Makefile| 1 +
libavfilter/allfilters.c| 1 +
libavfilter/vf_colorconstancy.c | 808
5 files changed, 812 insertions(+)
create mode 100644 libavfilter/vf_colorconstancy.c
diff --git a/Changelog b/Changelog
index e43a159b0d..dfdb2f38af 100644
--- a/Changelog
+++ b/Changelog
@@ -14,6 +14,7 @@ version :
- libtensorflow backend for DNN based filters like srcnn
- vc1 decoder is now bit-exact
- ATRAC9 decoder
+- colorconstancy filter
version 4.0:
diff --git a/MAINTAINERS b/MAINTAINERS
index 78f450dda6..234b655cef 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -332,6 +332,7 @@ Filters:
vf_bwdif Thomas Mundt (CC )
vf_chromakey.cTimo Rothenpieler
vf_colorchannelmixer.cPaul B Mahol
+ vf_colorconstancy.c Mina Sami(CC )
vf_colorbalance.c Paul B Mahol
vf_colorkey.c Timo Rothenpieler
vf_colorlevels.c Paul B Mahol
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 7735c26529..ebe43e1391 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -228,6 +228,7 @@ OBJS-$(CONFIG_FSPP_FILTER) += vf_fspp.o
OBJS-$(CONFIG_GBLUR_FILTER) += vf_gblur.o
OBJS-$(CONFIG_GEQ_FILTER)+= vf_geq.o
OBJS-$(CONFIG_GRADFUN_FILTER)+= vf_gradfun.o
+OBJS-$(CONFIG_GREYEDGE_FILTER) += vf_colorconstancy.o
OBJS-$(CONFIG_HALDCLUT_FILTER) += vf_lut3d.o framesync.o
OBJS-$(CONFIG_HFLIP_FILTER) += vf_hflip.o
OBJS-$(CONFIG_HISTEQ_FILTER) += vf_histeq.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index 0ded83ede2..9c99b03a81 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -217,6 +217,7 @@ extern AVFilter ff_vf_fspp;
extern AVFilter ff_vf_gblur;
extern AVFilter ff_vf_geq;
extern AVFilter ff_vf_gradfun;
+extern AVFilter ff_vf_greyedge;
extern AVFilter ff_vf_haldclut;
extern AVFilter ff_vf_hflip;
extern AVFilter ff_vf_histeq;
diff --git a/libavfilter/vf_colorconstancy.c b/libavfilter/vf_colorconstancy.c
new file mode 100644
index 00..ec7b2a1832
--- /dev/null
+++ b/libavfilter/vf_colorconstancy.c
@@ -0,0 +1,808 @@
+/*
+ * Copyright (c) 2018 Mina Sami
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file
+ * Color Constancy filter
+ *
+ * @see http://colorconstancy.com/
+ *
+ * @cite
+ * J. van de Weijer, Th. Gevers, A. Gijsenij "Edge-Based Color Constancy".
+ */
+
+#include "libavutil/avassert.h"
+#include "libavutil/bprint.h"
+#include "libavutil/imgutils.h"
+#include "libavutil/opt.h"
+#include "libavutil/pixdesc.h"
+
+#include "avfilter.h"
+#include "formats.h"
+#include "internal.h"
+#include "video.h"
+
+#include
+
+#define GREY_EDGE "greyedge"
+
+#define NUM_PLANES3
+#define MAX_DIFF_ORD 2
+#define MAX_META_DATA 4
+#define MAX_DATA 4
+
+#define INDEX_TEMP 0
+#define INDEX_DX 1
+#define INDEX_DY 2
+#define INDEX_DXY 3
+#define INDEX_NORM INDEX_DX
+#define INDEX_SRC 0
+#define INDEX_DST 1
+#define INDEX_ORD 2
+#define INDEX_DIR 3
+#define DIR_X 0
+#define DIR_Y 1
+
+/**
+ * Used for passing data between threads.
+ */
+typedef struct ThreadData {
+AVFrame *in, *out;
+int meta_data[MAX_META_DATA];
+double *data[MAX_DATA][NUM_PLANES];
+} ThreadData;
+
+/**
+ * Common struct for all algorithms contexts.
+ */
+typedef struct ColorConstancyContext {
+const AVClass *class;
+
+int correct;
+int save;
+char *outfile;
+
+int difford;
+int minknorm; /**< @minknorm = 0 : getMax instead */
+double sigm
[FFmpeg-devel] [PATCH v2][GSOC] avfilter: added colorconstancy
Hi,
This patch introduces Grey-Edge algorithm as part of the Color
Constancy Filter project in GSOC.
V2 changes:
- added documentation ini filters.texi
- remove currently not useful code (saving to file)
Best regards,
Mina Sami
>From deb1cad1bb007b7372dd128581033c5c63b63ee6 Mon Sep 17 00:00:00 2001
From: Mina
Date: Sun, 15 Jul 2018 04:16:13 +0200
Subject: [PATCH] avfilter: added colorconstancy
Signed-off-by: Mina
---
Changelog | 1 +
MAINTAINERS | 1 +
doc/filters.texi| 38 ++
libavfilter/Makefile| 1 +
libavfilter/allfilters.c| 1 +
libavfilter/vf_colorconstancy.c | 756
6 files changed, 798 insertions(+)
create mode 100644 libavfilter/vf_colorconstancy.c
diff --git a/Changelog b/Changelog
index e43a159b0d..dfdb2f38af 100644
--- a/Changelog
+++ b/Changelog
@@ -14,6 +14,7 @@ version :
- libtensorflow backend for DNN based filters like srcnn
- vc1 decoder is now bit-exact
- ATRAC9 decoder
+- colorconstancy filter
version 4.0:
diff --git a/MAINTAINERS b/MAINTAINERS
index 78f450dda6..234b655cef 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -332,6 +332,7 @@ Filters:
vf_bwdif Thomas Mundt (CC )
vf_chromakey.cTimo Rothenpieler
vf_colorchannelmixer.cPaul B Mahol
+ vf_colorconstancy.c Mina Sami(CC )
vf_colorbalance.c Paul B Mahol
vf_colorkey.c Timo Rothenpieler
vf_colorlevels.c Paul B Mahol
diff --git a/doc/filters.texi b/doc/filters.texi
index 49895ff947..a787068939 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -9940,6 +9940,44 @@ gradfun=radius=8
@end itemize
+@section greyedge
+
+A variation color constancy that estimates scene illumination via grey edge algorithm
+and corrects the scene colors accordingly.
+
+The filter accepts the following options:
+
+@table @option
+@item difford
+The order of diffrentation to be applied on the scene. Must be chosen in the range
+[0,2] and default value is 1.
+
+@item minknorm
+The Minkowski parameter to be used for calculating the Minkowski distance. Must
+be chosen in the range [0,65535] and default value is 1. Set to 0 for getting
+max value instead of calculating Minkowski distance.
+
+@item sigma
+The standard deviation of Gaussian blur to be applied on the scene. Must be
+chosen in the range [0,1024.0] and default value = 1. Sigma can't be set to 0
+if @var{difford} is greater than 1.
+
+@subsection Examples
+
+@item
+Grey Edge:
+@example
+greyedge=difford=1:minknorm=5:sigma=2
+@end example
+@end itemize
+
+@item
+Max Edge:
+@example
+greyedge=difford=1:minknorm=0:sigma=2
+@end example
+@end itemize
+
@anchor{haldclut}
@section haldclut
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 7735c26529..ebe43e1391 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -228,6 +228,7 @@ OBJS-$(CONFIG_FSPP_FILTER) += vf_fspp.o
OBJS-$(CONFIG_GBLUR_FILTER) += vf_gblur.o
OBJS-$(CONFIG_GEQ_FILTER)+= vf_geq.o
OBJS-$(CONFIG_GRADFUN_FILTER)+= vf_gradfun.o
+OBJS-$(CONFIG_GREYEDGE_FILTER) += vf_colorconstancy.o
OBJS-$(CONFIG_HALDCLUT_FILTER) += vf_lut3d.o framesync.o
OBJS-$(CONFIG_HFLIP_FILTER) += vf_hflip.o
OBJS-$(CONFIG_HISTEQ_FILTER) += vf_histeq.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index 0ded83ede2..9c99b03a81 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -217,6 +217,7 @@ extern AVFilter ff_vf_fspp;
extern AVFilter ff_vf_gblur;
extern AVFilter ff_vf_geq;
extern AVFilter ff_vf_gradfun;
+extern AVFilter ff_vf_greyedge;
extern AVFilter ff_vf_haldclut;
extern AVFilter ff_vf_hflip;
extern AVFilter ff_vf_histeq;
diff --git a/libavfilter/vf_colorconstancy.c b/libavfilter/vf_colorconstancy.c
new file mode 100644
index 00..281a4cb43d
--- /dev/null
+++ b/libavfilter/vf_colorconstancy.c
@@ -0,0 +1,756 @@
+/*
+ * Copyright (c) 2018 Mina Sami
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file
+
[FFmpeg-devel] Fwd: Re: [PATCH v2][GSOC] avfilter: added colorconstancy
Forwarded Message
Subject: Re: [FFmpeg-devel] [PATCH v2][GSOC] avfilter: added
colorconstancy
Date: Sun, 15 Jul 2018 20:51:09 +0200
From: Mina
To: Gyan Doshi
On 07/15/2018 07:16 AM, Gyan Doshi wrote:
On 15-07-2018 07:50 AM, Mina wrote:
First, thanks for your feedback.
+A variation color constancy that estimates scene illumination via
grey edge algorithm
+and corrects the scene colors accordingly.
"A variation color constancy filter which estimates ..."
Okay will be changed to "A color constancy variation filter which estimates"
Maybe add a link to the algo paper in the next line:
"Grey Edge algorithm: http://...";
The link I have is to a research paper so I didn't think a non-developer
user would benefit from it but will add it.
> +The order of diffrentation to be applied on the scene. Must be
chosen in the range
diffrentation -> differentiation
Changed.
> +chosen in the range [0,1024.0] and default value = 1. Sigma can't
be set to 0
> +if @var{difford} is greater than 1.
but for set_gauss, you say
> Sigma can be zero only at difford = 0
and test during config
> av_assert0(sigma || !s->difford);
Please correct the docs. Also prefer that you log an informational
error and return AVERROR(EINVAL) instead of an assert.
Sure, will do.
Thanks,
Gyan
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
[FFmpeg-devel] Fwd: Re: [PATCH v2][GSOC] avfilter: added colorconstancy
Forwarded Message Subject: Re: [FFmpeg-devel] [PATCH v2][GSOC] avfilter: added colorconstancy Date: Sun, 15 Jul 2018 20:51:51 +0200 From: Mina To: Reto Kromer On 07/15/2018 07:22 AM, Reto Kromer wrote: Hello Mina, I noticed a typo in the documentation: @item difford The order of diffrentation diffrentation -> differentiation Thanks for your feedback and it will be corrected. Best regards, Reto ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
[FFmpeg-devel] [PATCH v3][GSOC] avfilter: added colorconstancy
Hi,
This patch introduces Grey-Edge algorithm as part of the Color
Constancy Filter project in GSOC.
V3 changes:
- corrected typo in documentation
- corrected wrong value in documentation
- added link to algorithm research paper in documentation
- returned EINVAL and log error when sigma = 0 && difford > 0 instead of
assert
Best regards,
Mina Sami
>From 52a0057eda8bf5147b53c9d71089d944ebaa759c Mon Sep 17 00:00:00 2001
From: Mina
Date: Sun, 15 Jul 2018 20:59:52 +0200
Subject: [PATCH] avfilter: added colorconstancy
Signed-off-by: Mina
---
Changelog | 1 +
MAINTAINERS | 1 +
doc/filters.texi| 39 ++
libavfilter/Makefile| 1 +
libavfilter/allfilters.c| 1 +
libavfilter/vf_colorconstancy.c | 758
6 files changed, 801 insertions(+)
create mode 100644 libavfilter/vf_colorconstancy.c
diff --git a/Changelog b/Changelog
index 72da5bf519..807a05dec9 100644
--- a/Changelog
+++ b/Changelog
@@ -15,6 +15,7 @@ version :
- vc1 decoder is now bit-exact
- ATRAC9 decoder
- lensfun wrapper filter
+- colorconstancy filter
version 4.0:
diff --git a/MAINTAINERS b/MAINTAINERS
index 78f450dda6..234b655cef 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -332,6 +332,7 @@ Filters:
vf_bwdif Thomas Mundt (CC )
vf_chromakey.cTimo Rothenpieler
vf_colorchannelmixer.cPaul B Mahol
+ vf_colorconstancy.c Mina Sami(CC )
vf_colorbalance.c Paul B Mahol
vf_colorkey.c Timo Rothenpieler
vf_colorlevels.c Paul B Mahol
diff --git a/doc/filters.texi b/doc/filters.texi
index 9d8f88ddcf..a70f98ac6b 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -9940,6 +9940,45 @@ gradfun=radius=8
@end itemize
+@section greyedge
+
+A color constancy variation filter which estimates scene illumination via grey edge algorithm
+and corrects the scene colors accordingly.
+See: https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf
+
+The filter accepts the following options:
+
+@table @option
+@item difford
+The order of differentiation to be applied on the scene. Must be chosen in the range
+[0,2] and default value is 1.
+
+@item minknorm
+The Minkowski parameter to be used for calculating the Minkowski distance. Must
+be chosen in the range [0,65535] and default value is 1. Set to 0 for getting
+max value instead of calculating Minkowski distance.
+
+@item sigma
+The standard deviation of Gaussian blur to be applied on the scene. Must be
+chosen in the range [0,1024.0] and default value = 1. Sigma can't be set to 0
+if @var{difford} is greater than 0.
+
+@subsection Examples
+
+@item
+Grey Edge:
+@example
+greyedge=difford=1:minknorm=5:sigma=2
+@end example
+@end itemize
+
+@item
+Max Edge:
+@example
+greyedge=difford=1:minknorm=0:sigma=2
+@end example
+@end itemize
+
@anchor{haldclut}
@section haldclut
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 5d4549e24c..86b04e3efb 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -228,6 +228,7 @@ OBJS-$(CONFIG_FSPP_FILTER) += vf_fspp.o
OBJS-$(CONFIG_GBLUR_FILTER) += vf_gblur.o
OBJS-$(CONFIG_GEQ_FILTER)+= vf_geq.o
OBJS-$(CONFIG_GRADFUN_FILTER)+= vf_gradfun.o
+OBJS-$(CONFIG_GREYEDGE_FILTER) += vf_colorconstancy.o
OBJS-$(CONFIG_HALDCLUT_FILTER) += vf_lut3d.o framesync.o
OBJS-$(CONFIG_HFLIP_FILTER) += vf_hflip.o
OBJS-$(CONFIG_HISTEQ_FILTER) += vf_histeq.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index 521bc53164..2d19929bdc 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -217,6 +217,7 @@ extern AVFilter ff_vf_fspp;
extern AVFilter ff_vf_gblur;
extern AVFilter ff_vf_geq;
extern AVFilter ff_vf_gradfun;
+extern AVFilter ff_vf_greyedge;
extern AVFilter ff_vf_haldclut;
extern AVFilter ff_vf_hflip;
extern AVFilter ff_vf_histeq;
diff --git a/libavfilter/vf_colorconstancy.c b/libavfilter/vf_colorconstancy.c
new file mode 100644
index 00..7194688dfa
--- /dev/null
+++ b/libavfilter/vf_colorconstancy.c
@@ -0,0 +1,758 @@
+/*
+ * Copyright (c) 2018 Mina Sami
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General
Re: [FFmpeg-devel] [PATCH v3][GSOC] avfilter: added colorconstancy
On 07/16/2018 02:55 AM, Michael Niedermayer wrote: On Sun, Jul 15, 2018 at 09:13:04PM +0200, Mina wrote: Hi, This patch introduces Grey-Edge algorithm as part of the Color Constancy Filter project in GSOC. V3 changes: - corrected typo in documentation - corrected wrong value in documentation - added link to algorithm research paper in documentation - returned EINVAL and log error when sigma = 0 && difford > 0 instead of assert Best regards, Mina Sami Changelog |1 MAINTAINERS |1 doc/filters.texi| 39 ++ libavfilter/Makefile|1 libavfilter/allfilters.c|1 libavfilter/vf_colorconstancy.c | 758 6 files changed, 801 insertions(+) 470a6bcd3455fa6802ef290f438632a0b1e0ab6f 0001-avfilter-added-colorconstancy.patch From 52a0057eda8bf5147b53c9d71089d944ebaa759c Mon Sep 17 00:00:00 2001 From: Mina Date: Sun, 15 Jul 2018 20:59:52 +0200 Subject: [PATCH] avfilter: added colorconstancy breaks "make doc/ffprobe-all.html" HTMLdoc/ffprobe-all.html doc/filters.texi:9966: @subsection seen before @end table doc/filters.texi:9968: @item outside of table or list doc/filters.texi:9973: unmatched `@end itemize' doc/filters.texi:9975: @item outside of table or list doc/filters.texi:9980: unmatched `@end itemize' make: *** [doc/ffprobe-all.html] Error 1 [...] Hi, thanks for feedback. Fixed now. Any more edits to be added to v4? ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
[FFmpeg-devel] [PATCH v4][GSOC] avfilter: added colorconstancy
Hi,
This patch introduces Grey-Edge algorithm as part of the Color
Constancy Filter project in GSOC.
V4 changes:
- Fixed error in filter.texi that resulted in breaking "make
doc/ffprobe-all.html"
Best regards,
Mina Sami
>From 7e24e01bc3a8303e4876b6a455dfc7d933efb6aa Mon Sep 17 00:00:00 2001
From: Mina
Date: Mon, 16 Jul 2018 10:20:02 +0200
Subject: [PATCH] avfilter: added colorconstancy
Signed-off-by: Mina
---
Changelog | 1 +
MAINTAINERS | 1 +
doc/filters.texi| 41 ++
libavfilter/Makefile| 1 +
libavfilter/allfilters.c| 1 +
libavfilter/vf_colorconstancy.c | 758
6 files changed, 803 insertions(+)
create mode 100644 libavfilter/vf_colorconstancy.c
diff --git a/Changelog b/Changelog
index 72da5bf519..807a05dec9 100644
--- a/Changelog
+++ b/Changelog
@@ -15,6 +15,7 @@ version :
- vc1 decoder is now bit-exact
- ATRAC9 decoder
- lensfun wrapper filter
+- colorconstancy filter
version 4.0:
diff --git a/MAINTAINERS b/MAINTAINERS
index 78f450dda6..234b655cef 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -332,6 +332,7 @@ Filters:
vf_bwdif Thomas Mundt (CC )
vf_chromakey.cTimo Rothenpieler
vf_colorchannelmixer.cPaul B Mahol
+ vf_colorconstancy.c Mina Sami(CC )
vf_colorbalance.c Paul B Mahol
vf_colorkey.c Timo Rothenpieler
vf_colorlevels.c Paul B Mahol
diff --git a/doc/filters.texi b/doc/filters.texi
index 705d48e1b0..6e228f76ed 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -9940,6 +9940,47 @@ gradfun=radius=8
@end itemize
+@section greyedge
+A color constancy variation filter which estimates scene illumination via grey edge algorithm
+and corrects the scene colors accordingly.
+
+See: @url{https://staff.science.uva.nl/th.gevers/pub/GeversTIP07.pdf}
+
+The filter accepts the following options:
+
+@table @option
+@item difford
+The order of differentiation to be applied on the scene. Must be chosen in the range
+[0,2] and default value is 1.
+
+@item minknorm
+The Minkowski parameter to be used for calculating the Minkowski distance. Must
+be chosen in the range [0,65535] and default value is 1. Set to 0 for getting
+max value instead of calculating Minkowski distance.
+
+@item sigma
+The standard deviation of Gaussian blur to be applied on the scene. Must be
+chosen in the range [0,1024.0] and default value = 1. Sigma can't be set to 0
+if @var{difford} is greater than 0.
+@end table
+
+@subsection Examples
+@itemize
+
+@item
+Grey Edge:
+@example
+greyedge=difford=1:minknorm=5:sigma=2
+@end example
+
+@item
+Max Edge:
+@example
+greyedge=difford=1:minknorm=0:sigma=2
+@end example
+
+@end itemize
+
@anchor{haldclut}
@section haldclut
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 5d4549e24c..86b04e3efb 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -228,6 +228,7 @@ OBJS-$(CONFIG_FSPP_FILTER) += vf_fspp.o
OBJS-$(CONFIG_GBLUR_FILTER) += vf_gblur.o
OBJS-$(CONFIG_GEQ_FILTER)+= vf_geq.o
OBJS-$(CONFIG_GRADFUN_FILTER)+= vf_gradfun.o
+OBJS-$(CONFIG_GREYEDGE_FILTER) += vf_colorconstancy.o
OBJS-$(CONFIG_HALDCLUT_FILTER) += vf_lut3d.o framesync.o
OBJS-$(CONFIG_HFLIP_FILTER) += vf_hflip.o
OBJS-$(CONFIG_HISTEQ_FILTER) += vf_histeq.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index 521bc53164..2d19929bdc 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -217,6 +217,7 @@ extern AVFilter ff_vf_fspp;
extern AVFilter ff_vf_gblur;
extern AVFilter ff_vf_geq;
extern AVFilter ff_vf_gradfun;
+extern AVFilter ff_vf_greyedge;
extern AVFilter ff_vf_haldclut;
extern AVFilter ff_vf_hflip;
extern AVFilter ff_vf_histeq;
diff --git a/libavfilter/vf_colorconstancy.c b/libavfilter/vf_colorconstancy.c
new file mode 100644
index 00..7194688dfa
--- /dev/null
+++ b/libavfilter/vf_colorconstancy.c
@@ -0,0 +1,758 @@
+/*
+ * Copyright (c) 2018 Mina Sami
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin S
[FFmpeg-devel] [PATCH] lavfi: update xbr filter url
Hi,
While adding documentation in doc/filters.texi as part of [GSOC]Color
Constancy Filter project, I found the url for xbr filter giving an
internal server error. After some search I found another link where the
author of the filter(and algorithm) posts a tutorial for the algorithm.
This patch update the existing url with the newly found one.
Thanks,
Mina
>From 1193d48dfbb572ebbe5e3849950079f0af5765ed Mon Sep 17 00:00:00 2001
From: Mina
Date: Wed, 8 Aug 2018 21:43:18 +0200
Subject: [PATCH] lavfi: update xbr filter url
Signed-off-by: Mina
---
doc/filters.texi | 2 +-
libavfilter/vf_xbr.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/doc/filters.texi b/doc/filters.texi
index 0b0903e5a7..792edfd2e1 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -17148,7 +17148,7 @@ separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
@section xbr
Apply the xBR high-quality magnification filter which is designed for pixel
art. It follows a set of edge-detection rules, see
-@url{http://www.libretro.com/forums/viewtopic.php?f=6&t=134}.
+@url{https://forums.libretro.com/t/xbr-algorithm-tutorial/123}.
It accepts the following option:
diff --git a/libavfilter/vf_xbr.c b/libavfilter/vf_xbr.c
index 78094e0287..2c71871d22 100644
--- a/libavfilter/vf_xbr.c
+++ b/libavfilter/vf_xbr.c
@@ -24,7 +24,7 @@
* XBR Filter is used for depixelization of image.
* This is based on Hyllian's xBR shader.
*
- * @see http://www.libretro.com/forums/viewtopic.php?f=6&t=134
+ * @see https://forums.libretro.com/t/xbr-algorithm-tutorial/123
* @see https://github.com/yoyofr/iFBA/blob/master/fba_src/src/intf/video/scalers/xbr.cpp
*/
--
2.17.0
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
[FFmpeg-devel] [GSOC] [PATCH 1/2] lavfi/vf_colorconstancy: changing options ranges
Hi,
This patch changes the maximum value for minknorm to a more sane
value and fixes check for sigma value when difford > 0. It also
itroduces some minor cosmetic edits(spaces and so).
Thanks.
>From 8f4270e4aad285a8652e1831e828439cdb13620a Mon Sep 17 00:00:00 2001
From: Mina
Date: Mon, 13 Aug 2018 01:29:28 +0200
Subject: [PATCH 1/2] lavfi/vf_colorconstancy: changing options ranges
Signed-off-by: Mina
---
doc/filters.texi| 6 +++---
libavfilter/vf_colorconstancy.c | 20 ++--
2 files changed, 13 insertions(+), 13 deletions(-)
diff --git a/doc/filters.texi b/doc/filters.texi
index d6c15837f2..267bd04a43 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -9955,13 +9955,13 @@ The order of differentiation to be applied on the scene. Must be chosen in the r
@item minknorm
The Minkowski parameter to be used for calculating the Minkowski distance. Must
-be chosen in the range [0,65535] and default value is 1. Set to 0 for getting
+be chosen in the range [0,20] and default value is 1. Set to 0 for getting
max value instead of calculating Minkowski distance.
@item sigma
The standard deviation of Gaussian blur to be applied on the scene. Must be
-chosen in the range [0,1024.0] and default value = 1. Sigma can't be set to 0
-if @var{difford} is greater than 0.
+chosen in the range [0,1024.0] and default value = 1. floor( @var{sigma} * break_off_sigma(3) )
+can't be euqal to 0 if @var{difford} is greater than 0.
@end table
@subsection Examples
diff --git a/libavfilter/vf_colorconstancy.c b/libavfilter/vf_colorconstancy.c
index 7194688dfa..e3bb39e51b 100644
--- a/libavfilter/vf_colorconstancy.c
+++ b/libavfilter/vf_colorconstancy.c
@@ -28,7 +28,6 @@
* J. van de Weijer, Th. Gevers, A. Gijsenij "Edge-Based Color Constancy".
*/
-#include "libavutil/bprint.h"
#include "libavutil/imgutils.h"
#include "libavutil/opt.h"
#include "libavutil/pixdesc.h"
@@ -42,6 +41,8 @@
#define GREY_EDGE "greyedge"
+#define SQRT3 1.73205080757
+
#define NUM_PLANES3
#define MAX_DIFF_ORD 2
#define MAX_META_DATA 4
@@ -145,7 +146,7 @@ static int set_gauss(AVFilterContext *ctx)
sum1 = 0.0;
for (i = 0; i < filtersize; ++i) {
s->gauss[1][i] = - (GINDX(filtersize, i) / pow(sigma, 2)) * s->gauss[0][i];
-sum1 += s->gauss[1][i] *GINDX(filtersize, i);
+sum1 += s->gauss[1][i] * GINDX(filtersize, i);
}
for (i = 0; i < filtersize; ++i) {
@@ -595,7 +596,6 @@ static int diagonal_transformation(AVFilterContext *ctx, void *arg, int jobnr, i
ThreadData *td = arg;
AVFrame *in = td->in;
AVFrame *out = td->out;
-double sqrt3 = pow(3.0, 0.5);
int plane;
for (plane = 0; plane < NUM_PLANES; ++plane) {
@@ -610,7 +610,7 @@ static int diagonal_transformation(AVFilterContext *ctx, void *arg, int jobnr, i
unsigned i;
for (i = slice_start; i < slice_end; ++i) {
-temp = src[i] / (s->white[plane] * sqrt3);
+temp = src[i] / (s->white[plane] * SQRT3);
dst[i] = av_clip_uint8((int)(temp + 0.5));
}
}
@@ -657,12 +657,12 @@ static int config_props(AVFilterLink *inlink)
double sigma = s->sigma;
int ret;
-if (!sigma && s->difford) {
-av_log(ctx, AV_LOG_ERROR, "Sigma can't be set to 0 when difford > 0.\n");
+if (!floor(break_off_sigma * sigma + 0.5) && s->difford) {
+av_log(ctx, AV_LOG_ERROR, "floor(%f * sigma) must be > 0 when difford > 0.\n", break_off_sigma);
return AVERROR(EINVAL);
}
-s->filtersize = 2 * floor(break_off_sigma * s->sigma + 0.5) + 1;
+s->filtersize = 2 * floor(break_off_sigma * sigma + 0.5) + 1;
if (ret=set_gauss(ctx)) {
return ret;
}
@@ -735,9 +735,9 @@ static const AVFilterPad colorconstancy_outputs[] = {
#if CONFIG_GREYEDGE_FILTER
static const AVOption greyedge_options[] = {
-{ "difford", "set differentiation order", OFFSET(difford), AV_OPT_TYPE_INT,{.i64=1}, 0, 2, FLAGS },
-{ "minknorm", "set Minkowski norm",OFFSET(minknorm), AV_OPT_TYPE_INT,{.i64=1}, 0, 65535, FLAGS },
-{ "sigma","set sigma", OFFSET(sigma),AV_OPT_TYPE_DOUBLE, {.dbl=1}, 0.0, 1024.0, FLAGS },
+{ "difford", "set differentiation order", OFFSET(difford), AV_OPT_TYPE_INT,{.i64=1}, 0, 2, FLAGS },
+{ "minknorm", "set Minkowski norm",OFFSET(minknorm), AV_OPT_TYPE_INT,{.i64=1}, 0, 20, FLAGS },
+{ "sigma","set sigma", OFFSET(sigma),AV_OPT_TYPE_DOUBLE, {.dbl=1}, 0.0, 1024.0, FLAGS },
{ NULL }
};
--
2.17.0
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
[FFmpeg-devel] [GSOC] [PATCH 1/2] lavfi/vf_colorconstancy: adding weighted_ greyedge
Hi,
This patch introduces a new improved color constancy algorithm based
on previously implemented greyedge algorithm.
Thanks.
>From ec68204b9b081e0f0861ab50767c41559ed7e7b6 Mon Sep 17 00:00:00 2001
From: Mina
Date: Mon, 13 Aug 2018 01:49:24 +0200
Subject: [PATCH 2/2] lavfi/vf_colorconstancy: adding weighted_ greyedge
Signed-off-by: Mina
---
doc/filters.texi| 39
libavfilter/Makefile| 1 +
libavfilter/allfilters.c| 1 +
libavfilter/vf_colorconstancy.c | 322
4 files changed, 288 insertions(+), 75 deletions(-)
diff --git a/doc/filters.texi b/doc/filters.texi
index 267bd04a43..07dacec831 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -17154,6 +17154,45 @@ separatefields,select=eq(mod(n,4),0)+eq(mod(n,4),3),weave
@end example
@end itemize
+@section weighted_greyedge
+A color constancy variation filter which estimates scene illumination via weighted grey edge
+algorithm and corrects the scene colors accordingly.
+
+See: @url{http://refbase.cvc.uab.es/files/GGW2012.pdf}
+
+The filter accepts the following options:
+
+@table @option
+@item minknorm
+The Minkowski parameter to be used for calculating the Minkowski distance. Must
+be chosen in the range [1,20] and default value is 1.
+
+@item sigma
+The standard deviation of Gaussian blur to be applied on the scene. Must be
+chosen in the range [0.2,1024.0] and default value = 1.
+
+@item wmap_k
+The power applied to the weight map to emphasis heigher weights. Must be chosen
+in the range [1,20] and default value = 2.
+
+@item iters
+The max number of iterations for performing the algorithm. Must be chosen in the
+range [1,20] and default value = 5.
+
+@item angerr
+The angular error threshold in degrees for stoping the algorithm. Must be chosen
+in the range [0,360] and default value = 0.001.
+@end table
+
+@subsection Examples
+@itemize
+
+@item
+@example
+weighted_greyedge=minknorm=5:sigma=2:wmap_k=10:iters=10:angerr=0.0005
+@end example
+@end itemize
+
@section xbr
Apply the xBR high-quality magnification filter which is designed for pixel
art. It follows a set of edge-detection rules, see
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index e5d3a57af7..4086e941fa 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -392,6 +392,7 @@ OBJS-$(CONFIG_VSTACK_FILTER) += vf_stack.o framesync.o
OBJS-$(CONFIG_W3FDIF_FILTER) += vf_w3fdif.o
OBJS-$(CONFIG_WAVEFORM_FILTER) += vf_waveform.o
OBJS-$(CONFIG_WEAVE_FILTER) += vf_weave.o
+OBJS-$(CONFIG_WEIGHTED_GREYEDGE_FILTER) += vf_colorconstancy.o
OBJS-$(CONFIG_XBR_FILTER)+= vf_xbr.o
OBJS-$(CONFIG_YADIF_FILTER) += vf_yadif.o
OBJS-$(CONFIG_ZMQ_FILTER)+= f_zmq.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index 9732ae5345..8c088e2c1e 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -373,6 +373,7 @@ extern AVFilter ff_vf_vstack;
extern AVFilter ff_vf_w3fdif;
extern AVFilter ff_vf_waveform;
extern AVFilter ff_vf_weave;
+extern AVFilter ff_vf_weighted_greyedge;
extern AVFilter ff_vf_xbr;
extern AVFilter ff_vf_yadif;
extern AVFilter ff_vf_zmq;
diff --git a/libavfilter/vf_colorconstancy.c b/libavfilter/vf_colorconstancy.c
index e3bb39e51b..3b84a637c7 100644
--- a/libavfilter/vf_colorconstancy.c
+++ b/libavfilter/vf_colorconstancy.c
@@ -26,6 +26,14 @@
*
* @cite
* J. van de Weijer, Th. Gevers, A. Gijsenij "Edge-Based Color Constancy".
+ *
+ * @cite
+ * J. van de Weijer, Th. Gevers, and J. Geusebroek,
+ * âEdge and corner detection by photometric quasi-invariantsâ.
+ *
+ * @cite
+ * A. Gijsenij, Th. Gevers, J. van de Weijer,
+ * "Improving Color Constancy by Photometric Edge Weighting".
*/
#include "libavutil/imgutils.h"
@@ -39,7 +47,8 @@
#include
-#define GREY_EDGE "greyedge"
+#define GREY_EDGE "greyedge"
+#define WEIGHTED_GREY_EDGE "weighted_greyedge"
#define SQRT3 1.73205080757
@@ -79,6 +88,10 @@ typedef struct ColorConstancyContext {
int minknorm; /**< @minknorm = 0 : getMax instead */
double sigma;
+int wmap_k;
+int iters;
+double angerr_thresh;
+
int nb_threads;
int planeheight[4];
int planewidth[4];
@@ -477,53 +490,73 @@ static int filter_slice_grey_edge(AVFilterContext* ctx, void* arg, int jobnr, in
}
/**
- * Main control function for grey edge algorithm.
+ * Slice function for weighted grey edge algorithm that is called iteratively to
+ * calculate and apply weight scheme.
*
* @param ctx the filter context.
- * @param in frame to perfrom grey edge on.
+ * @param arg data to be passed between threads.
+ * @param jobnr current job nubmer.
+ * @param nb_jobs total number of jobs.
*
- * @return 0 in case of success, a negative value corresponding to an
- * AVERROR code in
Re: [FFmpeg-devel] [PATCH][GSOC] avfilter: added colorconstancy
On Fri, Jul 13, 2018 at 12:10 PM, Gyan Doshi wrote: > > > On 13-07-2018 02:44 PM, Mina wrote: > >> >>This patch introduces Grey-Edge algorithm as part of the Color >> Constancy Filter project in GSOC. >> > > Docs missing. > > The filter isn't completely finished yet and documentation is planed before the GSOC duration is finished. > Regards, > Gyan > ___ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel > ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
