[FFmpeg-devel] [PATCH] avfilter: add findandcover filter

Michael Niedermayer michaelni at gmx.at
Thu Apr 30 14:11:20 CEST 2015


Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
---
 configure                     |    1 +
 doc/filters.texi              |   46 +++++
 libavfilter/Makefile          |    1 +
 libavfilter/allfilters.c      |    1 +
 libavfilter/version.h         |    2 +-
 libavfilter/vf_findandcover.c |  399 +++++++++++++++++++++++++++++++++++++++++
 6 files changed, 449 insertions(+), 1 deletion(-)
 create mode 100644 libavfilter/vf_findandcover.c

diff --git a/configure b/configure
index 88e0d97..965a678 100755
--- a/configure
+++ b/configure
@@ -2644,6 +2644,7 @@ eq_filter_deps="gpl"
 fftfilt_filter_deps="avcodec"
 fftfilt_filter_select="rdft"
 flite_filter_deps="libflite"
+findandcover_filter_deps="gpl"
 frei0r_filter_deps="frei0r dlopen"
 frei0r_src_filter_deps="frei0r dlopen"
 fspp_filter_deps="gpl"
diff --git a/doc/filters.texi b/doc/filters.texi
index b7a9c9b..c5ddca1 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -5163,6 +5163,52 @@ framework.
 
 It does not take parameters.
 
+ at section foc
+
+Find and cover a rectangular object
+
+It accepts the following options:
+
+ at table @option
+ at item object
+Filepath of the object image, needs to be in gray8.
+
+ at item cover
+Filepath of the optional cover image, needs to be in yuv420.
+
+ at item threshold
+Detection threshold, default is 0.5.
+
+ at item mipmaps
+Number of mipmaps, default is 3.
+
+ at item xmin, ymin, xmax, ymax
+Specifies the recatangle in which to search.
+
+ at item mode
+Set covering mode.
+
+It accepts the following values:
+ at table @samp
+ at item cover
+cover it by the supplied image
+ at item blur
+cover it by interpolating the surrounding pixels
+ at end table
+
+Default value is @var{blur}.
+ at end table
+
+ at subsection Examples
+
+ at itemize
+ at item
+Generate a representative palette of a given video using @command{ffmpeg}:
+ at example
+ffmpeg -i file.ts -vf foc=newref.pgm:cover.jpg:mode=cover new.mkv
+ at end example
+ at end itemize
+
 @anchor{format}
 @section format
 
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 48cee50..09bc465 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -126,6 +126,7 @@ OBJS-$(CONFIG_FFTFILT_FILTER)                += vf_fftfilt.o
 OBJS-$(CONFIG_FIELD_FILTER)                  += vf_field.o
 OBJS-$(CONFIG_FIELDMATCH_FILTER)             += vf_fieldmatch.o
 OBJS-$(CONFIG_FIELDORDER_FILTER)             += vf_fieldorder.o
+OBJS-$(CONFIG_FINDANDCOVER_FILTER)           += vf_findandcover.o
 OBJS-$(CONFIG_FORMAT_FILTER)                 += vf_format.o
 OBJS-$(CONFIG_FRAMESTEP_FILTER)              += vf_framestep.o
 OBJS-$(CONFIG_FPS_FILTER)                    += vf_fps.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index 7961dca..c4d4d74 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -142,6 +142,7 @@ void avfilter_register_all(void)
     REGISTER_FILTER(FIELD,          field,          vf);
     REGISTER_FILTER(FIELDMATCH,     fieldmatch,     vf);
     REGISTER_FILTER(FIELDORDER,     fieldorder,     vf);
+    REGISTER_FILTER(FINDANDCOVER,   findandcover,   vf);
     REGISTER_FILTER(FORMAT,         format,         vf);
     REGISTER_FILTER(FPS,            fps,            vf);
     REGISTER_FILTER(FRAMEPACK,      framepack,      vf);
diff --git a/libavfilter/version.h b/libavfilter/version.h
index a4a9e63..d6dc95a 100644
--- a/libavfilter/version.h
+++ b/libavfilter/version.h
@@ -30,7 +30,7 @@
 #include "libavutil/version.h"
 
 #define LIBAVFILTER_VERSION_MAJOR  5
-#define LIBAVFILTER_VERSION_MINOR  14
+#define LIBAVFILTER_VERSION_MINOR  15
 #define LIBAVFILTER_VERSION_MICRO 100
 
 #define LIBAVFILTER_VERSION_INT AV_VERSION_INT(LIBAVFILTER_VERSION_MAJOR, \
diff --git a/libavfilter/vf_findandcover.c b/libavfilter/vf_findandcover.c
new file mode 100644
index 0000000..33947f0
--- /dev/null
+++ b/libavfilter/vf_findandcover.c
@@ -0,0 +1,399 @@
+/*
+ * Copyright (c) 2014-2015 Michael Niedermayer <michaelni at gmx.at>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 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 General Public License for more details.
+ *
+ * You should have received a copy of the GNU 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.
+ */
+
+#include "libavutil/avassert.h"
+#include "libavutil/imgutils.h"
+#include "libavutil/opt.h"
+#include "internal.h"
+
+#include "lavfutils.h"
+
+enum mode {
+    MODE_COVER,
+    MODE_BLUR,
+    NB_MODES
+};
+
+#define MAX_MIPMAPS 5
+
+typedef struct FOCContext {
+    AVClass *class;
+    float threshold;
+    int mipmaps;
+    int xmin, ymin, xmax, ymax;
+    int mode;
+    char *obj_filename;
+    char *cover_filename;
+    int last_x, last_y;
+    AVFrame *obj_frame;
+    AVFrame *cover_frame;
+    AVFrame *needle_frame[MAX_MIPMAPS];
+    AVFrame *haystack_frame[MAX_MIPMAPS];
+} FOCContext;
+
+#define OFFSET(x) offsetof(FOCContext, x)
+#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
+static const AVOption foc_options[] = {
+    { "object", "object bitmap filename", OFFSET(obj_filename), AV_OPT_TYPE_STRING, {.str = NULL}, .flags = FLAGS },
+    { "cover",  "cover bitmap filename",  OFFSET(cover_filename),  AV_OPT_TYPE_STRING, {.str = NULL}, .flags = FLAGS },
+    { "threshold", "set threshold", OFFSET(threshold), AV_OPT_TYPE_FLOAT, {.dbl = 0.5}, 0, 1.0, FLAGS },
+    { "mipmaps", "set mipmaps", OFFSET(mipmaps), AV_OPT_TYPE_INT, {.i64 = 3}, 1, MAX_MIPMAPS, FLAGS },
+    { "xmin", "", OFFSET(xmin), AV_OPT_TYPE_INT, {.i64 = 0}, 0, INT_MAX, FLAGS },
+    { "ymin", "", OFFSET(ymin), AV_OPT_TYPE_INT, {.i64 = 0}, 0, INT_MAX, FLAGS },
+    { "xmax", "", OFFSET(xmax), AV_OPT_TYPE_INT, {.i64 = 0}, 0, INT_MAX, FLAGS },
+    { "ymax", "", OFFSET(ymax), AV_OPT_TYPE_INT, {.i64 = 0}, 0, INT_MAX, FLAGS },
+    { "mode", "set removial mode", OFFSET(mode), AV_OPT_TYPE_INT, {.i64 = MODE_BLUR}, 0, NB_MODES - 1, FLAGS, "mode" },
+        { "cover", "cover area with bitmap", 0, AV_OPT_TYPE_CONST, {.i64 = MODE_COVER}, INT_MIN, INT_MAX, FLAGS, "mode" },
+        { "blur", "blur area", 0, AV_OPT_TYPE_CONST, {.i64 = MODE_BLUR}, INT_MIN, INT_MAX, FLAGS, "mode" },
+    { NULL }
+};
+
+static const AVClass foc_class = {
+    .class_name       = "foc",
+    .item_name        = av_default_item_name,
+    .option           = foc_options,
+    .version          = LIBAVUTIL_VERSION_INT,
+    .category         = AV_CLASS_CATEGORY_FILTER,
+};
+
+static int query_formats(AVFilterContext *ctx)
+{
+    static const enum PixelFormat pix_fmts[] = {
+        AV_PIX_FMT_YUV420P,
+        AV_PIX_FMT_YUVJ420P,
+        AV_PIX_FMT_NONE
+    };
+
+    return ff_set_common_formats(ctx, ff_make_format_list(pix_fmts));
+}
+
+static AVFrame *downscale(AVFrame *in)
+{
+    int x, y;
+    AVFrame *frame = av_frame_alloc();
+    uint8_t *src, *dst;
+    if (!frame)
+        return NULL;
+
+    frame->format = in->format;
+    frame->width  = (in->width + 1) / 2;
+    frame->height = (in->height+ 1) / 2;
+
+    if (av_frame_get_buffer(frame, 32) < 0) {
+        av_frame_free(&frame);
+        return NULL;
+    }
+    src = in   ->data[0];
+    dst = frame->data[0];
+
+    for(y = 0; y<frame->height; y++) {
+        for(x = 0; x<frame->width; x++) {
+            dst[x] = (  src[2*x+0]
+                      + src[2*x+1]
+                      + src[2*x+0 + in->linesize[0]]
+                      + src[2*x+1 + in->linesize[0]]
+                      + 2) >> 2;
+        }
+        src += 2*in->linesize[0];
+        dst += frame->linesize[0];
+    }
+    return frame;
+}
+
+static float compare(const AVFrame *haystack, const AVFrame *obj, int offx, int offy)
+{
+    int x,y;
+    int o_sum_v = 0;
+    int h_sum_v = 0;
+    int64_t oo_sum_v = 0;
+    int64_t hh_sum_v = 0;
+    int64_t oh_sum_v = 0;
+    float c;
+    int n = obj->height * obj->width;
+    const uint8_t *odat = obj     ->data[0];
+    const uint8_t *hdat = haystack->data[0] + offx + offy * haystack->linesize[0];
+    int64_t o_sigma, h_sigma;
+
+    for(y = 0; y<obj->height; y++) {
+        for(x = 0; x<obj->width; x++) {
+            int o_v = odat[x];
+            int h_v = hdat[x];
+            o_sum_v += o_v;
+            h_sum_v += h_v;
+            oo_sum_v += o_v * o_v;
+            hh_sum_v += h_v * h_v;
+            oh_sum_v += o_v * h_v;
+        }
+        odat += obj->linesize[0];
+        hdat += haystack->linesize[0];
+    }
+    o_sigma = n*oo_sum_v - o_sum_v*(int64_t)o_sum_v;
+    h_sigma = n*hh_sum_v - h_sum_v*(int64_t)h_sum_v;
+
+    if (o_sigma == 0 || h_sigma == 0)
+        return 1.0;
+
+    c = (n*oh_sum_v - o_sum_v*(int64_t)h_sum_v) / (sqrt(o_sigma)*sqrt(h_sigma));
+
+    return 1 - fabs(c);
+}
+
+static int config_input(AVFilterLink *inlink)
+{
+    AVFilterContext *ctx = inlink->dst;
+    FOCContext *foc = ctx->priv;
+
+    if (foc->xmax <= 0)
+        foc->xmax = inlink->w - foc->obj_frame->width;
+    if (foc->ymax <= 0)
+        foc->ymax = inlink->h - foc->obj_frame->height;
+
+    return 0;
+}
+
+static float search(FOCContext *foc, int pass, int maxpass, int xmin, int xmax, int ymin, int ymax, int *best_x, int *best_y, float best_score)
+{
+    int x, y;
+
+    if (pass + 1 <= maxpass) {
+        int sub_x, sub_y;
+        search(foc, pass+1, maxpass, xmin>>1, (xmax+1)>>1, ymin>>1, (ymax+1)>>1, &sub_x, &sub_y, 1.0);
+        xmin = FFMAX(xmin, 2*sub_x - 4);
+        xmax = FFMIN(xmax, 2*sub_x + 4);
+        ymin = FFMAX(ymin, 2*sub_y - 4);
+        ymax = FFMIN(ymax, 2*sub_y + 4);
+    }
+
+    for (y=ymin; y<=ymax; y++) {
+        for (x=xmin; x<=xmax; x++) {
+            float score = compare(foc->haystack_frame[pass], foc->needle_frame[pass], x, y);
+            av_assert0(score != 0);
+            if (score < best_score) {
+                best_score = score;
+                *best_x = x;
+                *best_y = y;
+            }
+        }
+    }
+    return best_score;
+}
+
+static void cover(FOCContext *foc, AVFrame *in, int offx, int offy)
+{
+    int x, y, p;
+
+    for (p=0; p<3; p++) {
+        uint8_t *data = in->data[p] + (offx>>!!p) + (offy>>!!p) * in->linesize[p];
+        const uint8_t *src = foc->cover_frame->data[p];
+        int w = FF_CEIL_RSHIFT(foc->obj_frame->width , !!p);
+        int h = FF_CEIL_RSHIFT(foc->obj_frame->height, !!p);
+        for (y=0; y<h; y++) {
+            for (x=0; x<w; x++) {
+                data[x] = src[x];
+            }
+            data += in->linesize[p];
+            src += foc->cover_frame->linesize[p];
+        }
+    }
+}
+static void blur(FOCContext *foc, AVFrame *in, int offx, int offy)
+{
+    int x, y, p;
+
+    for (p=0; p<3; p++) {
+        int ox = offx>>!!p;
+        int oy = offy>>!!p;
+        int stride = in->linesize[p];
+        uint8_t *data = in->data[p] + ox + oy * stride;
+        int w = FF_CEIL_RSHIFT(foc->obj_frame->width , !!p);
+        int h = FF_CEIL_RSHIFT(foc->obj_frame->height, !!p);
+        int iw = FF_CEIL_RSHIFT(in->width , !!p);
+        int ih = FF_CEIL_RSHIFT(in->height, !!p);
+        for (y=0; y<h; y++) {
+            for (x=0; x<w; x++) {
+                int c = 0;
+                int s = 0;
+                if (ox) {
+                    int scale = 65536 / (x + 1);
+                    s += data[-1 + y*stride] * scale;
+                    c += scale;
+                }
+                if (oy) {
+                    int scale = 65536 / (y + 1);
+                    s += data[x - stride] * scale;
+                    c += scale;
+                }
+                if (ox + w < iw) {
+                    int scale = 65536 / (w - x);
+                    s += data[w + y*stride] * scale;
+                    c += scale;
+                }
+                if (oy + h < ih) {
+                    int scale = 65536 / (h - y);
+                    s += data[x + h*stride] * scale;
+                    c += scale;
+                }
+                data[x + y*stride] = (s + (c>>1)) / c;
+            }
+        }
+    }
+}
+
+static int filter_frame(AVFilterLink *inlink, AVFrame *in)
+{
+    AVFilterContext *ctx = inlink->dst;
+    FOCContext *foc = ctx->priv;
+    float best_score;
+    int best_x, best_y;
+    int i;
+
+    foc->haystack_frame[0] = av_frame_clone(in);
+    for (i=1; i<foc->mipmaps; i++) {
+        foc->haystack_frame[i] = downscale(foc->haystack_frame[i-1]);
+    }
+
+    best_score = search(foc, 0, 0,
+                        FFMAX(foc->xmin, foc->last_x - 8),
+                        FFMIN(foc->xmax, foc->last_x + 8),
+                        FFMAX(foc->ymin, foc->last_y - 8),
+                        FFMIN(foc->ymax, foc->last_y + 8),
+                        &best_x, &best_y, 1.0);
+
+    best_score = search(foc, 0, foc->mipmaps - 1, foc->xmin, foc->xmax, foc->ymin, foc->ymax,
+                        &best_x, &best_y, best_score);
+
+    for (i=0; i<MAX_MIPMAPS; i++) {
+        av_frame_free(&foc->haystack_frame[i]);
+    }
+
+    if (best_score > foc->threshold) {
+        return ff_filter_frame(ctx->outputs[0], in);
+    }
+
+    av_log(ctx, AV_LOG_DEBUG, "Found at %d %d score %f\n", best_x, best_y, best_score);
+    foc->last_x = best_x;
+    foc->last_y = best_y;
+
+    av_frame_make_writable(in);
+
+    if (foc->mode == MODE_BLUR) {
+        blur (foc, in, best_x, best_y);
+    } else {
+        cover(foc, in, best_x, best_y);
+    }
+    return ff_filter_frame(ctx->outputs[0], in);
+}
+
+static av_cold void uninit(AVFilterContext *ctx)
+{
+    FOCContext *foc = ctx->priv;
+    int i;
+
+    for (i=0; i<MAX_MIPMAPS; i++) {
+        av_frame_free(&foc->needle_frame[i]);
+        av_frame_free(&foc->haystack_frame[i]);
+    }
+
+    if (foc->obj_frame)
+        av_freep(&foc->obj_frame->data[0]);
+    if (foc->cover_frame)
+        av_freep(&foc->cover_frame->data[0]);
+    av_frame_free(&foc->obj_frame);
+}
+
+static av_cold int init(AVFilterContext *ctx)
+{
+    FOCContext *foc = ctx->priv;
+    int ret, i;
+
+    if (!foc->obj_filename || (!foc->cover_filename && foc->mode == MODE_COVER)) {
+        av_log(ctx, AV_LOG_ERROR, "object or cover filename not set\n");
+        return AVERROR(EINVAL);
+    }
+
+    foc->obj_frame = av_frame_alloc();
+    if (!foc->obj_frame)
+        return AVERROR(ENOMEM);
+
+    if ((ret = ff_load_image(foc->obj_frame->data, foc->obj_frame->linesize,
+                             &foc->obj_frame->width, &foc->obj_frame->height,
+                             &foc->obj_frame->format, foc->obj_filename, ctx)) < 0)
+        return ret;
+
+    if (foc->obj_frame->format != AV_PIX_FMT_GRAY8) {
+        av_log(ctx, AV_LOG_ERROR, "object image is not a grayscale image\n");
+        return AVERROR(EINVAL);
+    }
+
+    foc->needle_frame[0] = av_frame_clone(foc->obj_frame);
+    for (i=1; i<foc->mipmaps; i++) {
+        foc->needle_frame[i] = downscale(foc->needle_frame[i-1]);
+        if (!foc->needle_frame[i])
+            return AVERROR(ENOMEM);
+    }
+
+    if (foc->mode == MODE_COVER) {
+        foc->cover_frame = av_frame_alloc();
+        if (!foc->cover_frame)
+            return AVERROR(ENOMEM);
+
+        if ((ret = ff_load_image(foc->cover_frame->data, foc->cover_frame->linesize,
+                                &foc->cover_frame->width, &foc->cover_frame->height,
+                                &foc->cover_frame->format, foc->cover_filename, ctx)) < 0)
+            return ret;
+
+        if (foc->cover_frame->format != AV_PIX_FMT_YUV420P && foc->cover_frame->format != AV_PIX_FMT_YUVJ420P) {
+            av_log(ctx, AV_LOG_ERROR, "cover image is not a YUV420 image\n");
+            return AVERROR(EINVAL);
+        }
+    }
+
+    return 0;
+}
+
+static const AVFilterPad foc_inputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_VIDEO,
+        .config_props = config_input,
+        .filter_frame = filter_frame,
+    },
+    { NULL }
+};
+
+static const AVFilterPad foc_outputs[] = {
+    {
+        .name = "default",
+        .type = AVMEDIA_TYPE_VIDEO,
+    },
+    { NULL }
+};
+
+AVFilter ff_vf_findandcover = {
+    .name            = "foc",
+    .description     = NULL_IF_CONFIG_SMALL("Find and cover a user specified object"),
+    .priv_size       = sizeof(FOCContext),
+    .init            = init,
+    .uninit          = uninit,
+    .query_formats   = query_formats,
+    .inputs          = foc_inputs,
+    .outputs         = foc_outputs,
+    .priv_class      = &foc_class,
+};
-- 
1.7.9.5



More information about the ffmpeg-devel mailing list