[FFmpeg-devel] [PATCH V2 6/6] lavfi/dnn_classify: add filter dnn_classify for classification based on detection bounding boxes

Guo, Yejun yejun.guo at intel.com
Thu Apr 29 16:36:57 EEST 2021


classification is done on every detection bounding box in frame's side data,
which are the results of object detection (filter dnn_detect).

Please refer to commit log of dnn_detect for the material for detection,
and see below for classification.

- download material for classifcation:
wget https://github.com/guoyejun/ffmpeg_dnn/raw/main/models/openvino/2021.1/emotions-recognition-retail-0003.bin
wget https://github.com/guoyejun/ffmpeg_dnn/raw/main/models/openvino/2021.1/emotions-recognition-retail-0003.xml
wget https://github.com/guoyejun/ffmpeg_dnn/raw/main/models/openvino/2021.1/emotions-recognition-retail-0003.label

- run command as:
./ffmpeg -i cici.jpg -vf dnn_detect=dnn_backend=openvino:model=face-detection-adas-0001.xml:input=data:output=detection_out:confidence=0.6:labels=face-detection-adas-0001.label,dnn_classify=dnn_backend=openvino:model=emotions-recognition-retail-0003.xml:input=data:output=prob_emotion:confidence=0.3:labels=emotions-recognition-retail-0003.label:target=face,showinfo -f null -

We'll see the detect&classify result as below:
[Parsed_showinfo_2 @ 0x55b7d25e77c0]   side data - detection bounding boxes:
[Parsed_showinfo_2 @ 0x55b7d25e77c0] source: face-detection-adas-0001.xml, emotions-recognition-retail-0003.xml
[Parsed_showinfo_2 @ 0x55b7d25e77c0] index: 0,  region: (1005, 813) -> (1086, 905), label: face, confidence: 10000/10000.
[Parsed_showinfo_2 @ 0x55b7d25e77c0]            classify:  label: happy, confidence: 6757/10000.
[Parsed_showinfo_2 @ 0x55b7d25e77c0] index: 1,  region: (888, 839) -> (967, 926), label: face, confidence: 6917/10000.
[Parsed_showinfo_2 @ 0x55b7d25e77c0]            classify:  label: anger, confidence: 4320/10000.

Signed-off-by: Guo, Yejun <yejun.guo at intel.com>
---
the main change of V2 in this patch set is to rebase with latest code
by resolving the conflicts.

 configure                     |   1 +
 doc/filters.texi              |  39 ++++
 libavfilter/Makefile          |   1 +
 libavfilter/allfilters.c      |   1 +
 libavfilter/vf_dnn_classify.c | 330 ++++++++++++++++++++++++++++++++++
 5 files changed, 372 insertions(+)
 create mode 100644 libavfilter/vf_dnn_classify.c

diff --git a/configure b/configure
index 820f719a32..9f2dfaf2d4 100755
--- a/configure
+++ b/configure
@@ -3550,6 +3550,7 @@ derain_filter_select="dnn"
 deshake_filter_select="pixelutils"
 deshake_opencl_filter_deps="opencl"
 dilation_opencl_filter_deps="opencl"
+dnn_classify_filter_select="dnn"
 dnn_detect_filter_select="dnn"
 dnn_processing_filter_select="dnn"
 drawtext_filter_deps="libfreetype"
diff --git a/doc/filters.texi b/doc/filters.texi
index 36e35a175b..b405cc5dfb 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -10127,6 +10127,45 @@ ffmpeg -i INPUT -f lavfi -i nullsrc=hd720,geq='r=128+80*(sin(sqrt((X-W/2)*(X-W/2
 @end example
 @end itemize
 
+ at section dnn_classify
+
+Do classification with deep neural networks based on bounding boxes.
+
+The filter accepts the following options:
+
+ at table @option
+ at item dnn_backend
+Specify which DNN backend to use for model loading and execution. This option accepts
+only openvino now, tensorflow backends will be added.
+
+ at item model
+Set path to model file specifying network architecture and its parameters.
+Note that different backends use different file formats.
+
+ at item input
+Set the input name of the dnn network.
+
+ at item output
+Set the output name of the dnn network.
+
+ at item confidence
+Set the confidence threshold (default: 0.5).
+
+ at item labels
+Set path to label file specifying the mapping between label id and name.
+Each label name is written in one line, tailing spaces and empty lines are skipped.
+The first line is the name of label id 0,
+and the second line is the name of label id 1, etc.
+The label id is considered as name if the label file is not provided.
+
+ at item backend_configs
+Set the configs to be passed into backend
+
+For tensorflow backend, you can set its configs with @option{sess_config} options,
+please use tools/python/tf_sess_config.py to get the configs for your system.
+
+ at end table
+
 @section dnn_detect
 
 Do object detection with deep neural networks.
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 5a287364b0..6c22d0404e 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -243,6 +243,7 @@ OBJS-$(CONFIG_DILATION_FILTER)               += vf_neighbor.o
 OBJS-$(CONFIG_DILATION_OPENCL_FILTER)        += vf_neighbor_opencl.o opencl.o \
                                                 opencl/neighbor.o
 OBJS-$(CONFIG_DISPLACE_FILTER)               += vf_displace.o framesync.o
+OBJS-$(CONFIG_DNN_CLASSIFY_FILTER)           += vf_dnn_classify.o
 OBJS-$(CONFIG_DNN_DETECT_FILTER)             += vf_dnn_detect.o
 OBJS-$(CONFIG_DNN_PROCESSING_FILTER)         += vf_dnn_processing.o
 OBJS-$(CONFIG_DOUBLEWEAVE_FILTER)            += vf_weave.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index 931d7dbb0d..87c3661cf4 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -229,6 +229,7 @@ extern const AVFilter ff_vf_detelecine;
 extern const AVFilter ff_vf_dilation;
 extern const AVFilter ff_vf_dilation_opencl;
 extern const AVFilter ff_vf_displace;
+extern const AVFilter ff_vf_dnn_classify;
 extern const AVFilter ff_vf_dnn_detect;
 extern const AVFilter ff_vf_dnn_processing;
 extern const AVFilter ff_vf_doubleweave;
diff --git a/libavfilter/vf_dnn_classify.c b/libavfilter/vf_dnn_classify.c
new file mode 100644
index 0000000000..18fcd452d0
--- /dev/null
+++ b/libavfilter/vf_dnn_classify.c
@@ -0,0 +1,330 @@
+/*
+ * 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
+ * implementing an classification filter using deep learning networks.
+ */
+
+#include "libavformat/avio.h"
+#include "libavutil/opt.h"
+#include "libavutil/pixdesc.h"
+#include "libavutil/avassert.h"
+#include "libavutil/imgutils.h"
+#include "filters.h"
+#include "dnn_filter_common.h"
+#include "formats.h"
+#include "internal.h"
+#include "libavutil/time.h"
+#include "libavutil/avstring.h"
+#include "libavutil/detection_bbox.h"
+
+typedef struct DnnClassifyContext {
+    const AVClass *class;
+    DnnContext dnnctx;
+    float confidence;
+    char *labels_filename;
+    char *target;
+    char **labels;
+    int label_count;
+} DnnClassifyContext;
+
+#define OFFSET(x) offsetof(DnnClassifyContext, dnnctx.x)
+#define OFFSET2(x) offsetof(DnnClassifyContext, x)
+#define FLAGS AV_OPT_FLAG_FILTERING_PARAM | AV_OPT_FLAG_VIDEO_PARAM
+static const AVOption dnn_classify_options[] = {
+    { "dnn_backend", "DNN backend",                OFFSET(backend_type),     AV_OPT_TYPE_INT,       { .i64 = 2 },    INT_MIN, INT_MAX, FLAGS, "backend" },
+#if (CONFIG_LIBOPENVINO == 1)
+    { "openvino",    "openvino backend flag",      0,                        AV_OPT_TYPE_CONST,     { .i64 = 2 },    0, 0, FLAGS, "backend" },
+#endif
+    DNN_COMMON_OPTIONS
+    { "confidence",  "threshold of confidence",    OFFSET2(confidence),      AV_OPT_TYPE_FLOAT,     { .dbl = 0.5 },  0, 1, FLAGS},
+    { "labels",      "path to labels file",        OFFSET2(labels_filename), AV_OPT_TYPE_STRING,    { .str = NULL }, 0, 0, FLAGS },
+    { "target",      "which one to be classified", OFFSET2(target),          AV_OPT_TYPE_STRING,    { .str = NULL }, 0, 0, FLAGS },
+    { NULL }
+};
+
+AVFILTER_DEFINE_CLASS(dnn_classify);
+
+static int dnn_classify_post_proc(AVFrame *frame, DNNData *output, uint32_t bbox_index, AVFilterContext *filter_ctx)
+{
+    DnnClassifyContext *ctx = filter_ctx->priv;
+    float conf_threshold = ctx->confidence;
+    AVDetectionBBoxHeader *header;
+    AVDetectionBBox *bbox;
+    float *classifications;
+    uint32_t label_id;
+    float confidence;
+    AVFrameSideData *sd;
+
+    if (output->channels <= 0) {
+        return -1;
+    }
+
+    sd = av_frame_get_side_data(frame, AV_FRAME_DATA_DETECTION_BBOXES);
+    header = (AVDetectionBBoxHeader *)sd->data;
+
+    if (bbox_index == 0) {
+        av_strlcat(header->source, ", ", sizeof(header->source));
+        av_strlcat(header->source, ctx->dnnctx.model_filename, sizeof(header->source));
+    }
+
+    classifications = output->data;
+    label_id = 0;
+    confidence= classifications[0];
+    for (int i = 1; i < output->channels; i++) {
+        if (classifications[i] > confidence) {
+            label_id = i;
+            confidence= classifications[i];
+        }
+    }
+
+    if (confidence < conf_threshold) {
+        return 0;
+    }
+
+    bbox = av_get_detection_bbox(header, bbox_index);
+    bbox->classify_confidences[bbox->classify_count] = av_make_q((int)(confidence * 10000), 10000);
+
+    if (ctx->labels && label_id < ctx->label_count) {
+        av_strlcpy(bbox->classify_labels[bbox->classify_count], ctx->labels[label_id], sizeof(bbox->classify_labels[bbox->classify_count]));
+    } else {
+        snprintf(bbox->classify_labels[bbox->classify_count], sizeof(bbox->classify_labels[bbox->classify_count]), "%d", label_id);
+    }
+
+    bbox->classify_count++;
+
+    return 0;
+}
+
+static void free_classify_labels(DnnClassifyContext *ctx)
+{
+    for (int i = 0; i < ctx->label_count; i++) {
+        av_freep(&ctx->labels[i]);
+    }
+    ctx->label_count = 0;
+    av_freep(&ctx->labels);
+}
+
+static int read_classify_label_file(AVFilterContext *context)
+{
+    int line_len;
+    FILE *file;
+    DnnClassifyContext *ctx = context->priv;
+
+    file = av_fopen_utf8(ctx->labels_filename, "r");
+    if (!file){
+        av_log(context, AV_LOG_ERROR, "failed to open file %s\n", ctx->labels_filename);
+        return AVERROR(EINVAL);
+    }
+
+    while (!feof(file)) {
+        char *label;
+        char buf[256];
+        if (!fgets(buf, 256, file)) {
+            break;
+        }
+
+        line_len = strlen(buf);
+        while (line_len) {
+            int i = line_len - 1;
+            if (buf[i] == '\n' || buf[i] == '\r' || buf[i] == ' ') {
+                buf[i] = '\0';
+                line_len--;
+            } else {
+                break;
+            }
+        }
+
+        if (line_len == 0)  // empty line
+            continue;
+
+        if (line_len >= AV_DETECTION_BBOX_LABEL_NAME_MAX_SIZE) {
+            av_log(context, AV_LOG_ERROR, "label %s too long\n", buf);
+            fclose(file);
+            return AVERROR(EINVAL);
+        }
+
+        label = av_strdup(buf);
+        if (!label) {
+            av_log(context, AV_LOG_ERROR, "failed to allocate memory for label %s\n", buf);
+            fclose(file);
+            return AVERROR(ENOMEM);
+        }
+
+        if (av_dynarray_add_nofree(&ctx->labels, &ctx->label_count, label) < 0) {
+            av_log(context, AV_LOG_ERROR, "failed to do av_dynarray_add\n");
+            fclose(file);
+            av_freep(&label);
+            return AVERROR(ENOMEM);
+        }
+    }
+
+    fclose(file);
+    return 0;
+}
+
+static av_cold int dnn_classify_init(AVFilterContext *context)
+{
+    DnnClassifyContext *ctx = context->priv;
+    int ret = ff_dnn_init(&ctx->dnnctx, DFT_ANALYTICS_CLASSIFY, context);
+    if (ret < 0)
+        return ret;
+    ff_dnn_set_classify_post_proc(&ctx->dnnctx, dnn_classify_post_proc);
+
+    if (ctx->labels_filename) {
+        return read_classify_label_file(context);
+    }
+    return 0;
+}
+
+static int dnn_classify_query_formats(AVFilterContext *context)
+{
+    static const enum AVPixelFormat pix_fmts[] = {
+        AV_PIX_FMT_RGB24, AV_PIX_FMT_BGR24,
+        AV_PIX_FMT_GRAY8, AV_PIX_FMT_GRAYF32,
+        AV_PIX_FMT_YUV420P, AV_PIX_FMT_YUV422P,
+        AV_PIX_FMT_YUV444P, AV_PIX_FMT_YUV410P, AV_PIX_FMT_YUV411P,
+        AV_PIX_FMT_NV12,
+        AV_PIX_FMT_NONE
+    };
+    AVFilterFormats *fmts_list = ff_make_format_list(pix_fmts);
+    return ff_set_common_formats(context, fmts_list);
+}
+
+static int dnn_classify_flush_frame(AVFilterLink *outlink, int64_t pts, int64_t *out_pts)
+{
+    DnnClassifyContext *ctx = outlink->src->priv;
+    int ret;
+    DNNAsyncStatusType async_state;
+
+    ret = ff_dnn_flush(&ctx->dnnctx);
+    if (ret != DNN_SUCCESS) {
+        return -1;
+    }
+
+    do {
+        AVFrame *in_frame = NULL;
+        AVFrame *out_frame = NULL;
+        async_state = ff_dnn_get_async_result(&ctx->dnnctx, &in_frame, &out_frame);
+        if (out_frame) {
+            av_assert0(in_frame == out_frame);
+            ret = ff_filter_frame(outlink, out_frame);
+            if (ret < 0)
+                return ret;
+            if (out_pts)
+                *out_pts = out_frame->pts + pts;
+        }
+        av_usleep(5000);
+    } while (async_state >= DAST_NOT_READY);
+
+    return 0;
+}
+
+static int dnn_classify_activate(AVFilterContext *filter_ctx)
+{
+    AVFilterLink *inlink = filter_ctx->inputs[0];
+    AVFilterLink *outlink = filter_ctx->outputs[0];
+    DnnClassifyContext *ctx = filter_ctx->priv;
+    AVFrame *in = NULL;
+    int64_t pts;
+    int ret, status;
+    int got_frame = 0;
+    int async_state;
+
+    FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink);
+
+    do {
+        // drain all input frames
+        ret = ff_inlink_consume_frame(inlink, &in);
+        if (ret < 0)
+            return ret;
+        if (ret > 0) {
+            if (ff_dnn_execute_model_classification(&ctx->dnnctx, in, in, ctx->target) != DNN_SUCCESS) {
+                return AVERROR(EIO);
+            }
+        }
+    } while (ret > 0);
+
+    // drain all processed frames
+    do {
+        AVFrame *in_frame = NULL;
+        AVFrame *out_frame = NULL;
+        async_state = ff_dnn_get_async_result(&ctx->dnnctx, &in_frame, &out_frame);
+        if (out_frame) {
+            av_assert0(in_frame == out_frame);
+            ret = ff_filter_frame(outlink, out_frame);
+            if (ret < 0)
+                return ret;
+            got_frame = 1;
+        }
+    } while (async_state == DAST_SUCCESS);
+
+    // if frame got, schedule to next filter
+    if (got_frame)
+        return 0;
+
+    if (ff_inlink_acknowledge_status(inlink, &status, &pts)) {
+        if (status == AVERROR_EOF) {
+            int64_t out_pts = pts;
+            ret = dnn_classify_flush_frame(outlink, pts, &out_pts);
+            ff_outlink_set_status(outlink, status, out_pts);
+            return ret;
+        }
+    }
+
+    FF_FILTER_FORWARD_WANTED(outlink, inlink);
+
+    return 0;
+}
+
+static av_cold void dnn_classify_uninit(AVFilterContext *context)
+{
+    DnnClassifyContext *ctx = context->priv;
+    ff_dnn_uninit(&ctx->dnnctx);
+    free_classify_labels(ctx);
+}
+
+static const AVFilterPad dnn_classify_inputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_VIDEO,
+    },
+    { NULL }
+};
+
+static const AVFilterPad dnn_classify_outputs[] = {
+    {
+        .name = "default",
+        .type = AVMEDIA_TYPE_VIDEO,
+    },
+    { NULL }
+};
+
+const AVFilter ff_vf_dnn_classify = {
+    .name          = "dnn_classify",
+    .description   = NULL_IF_CONFIG_SMALL("Apply DNN classify filter to the input."),
+    .priv_size     = sizeof(DnnClassifyContext),
+    .init          = dnn_classify_init,
+    .uninit        = dnn_classify_uninit,
+    .query_formats = dnn_classify_query_formats,
+    .inputs        = dnn_classify_inputs,
+    .outputs       = dnn_classify_outputs,
+    .priv_class    = &dnn_classify_class,
+    .activate      = dnn_classify_activate,
+};
-- 
2.17.1



More information about the ffmpeg-devel mailing list