[FFmpeg-devel,V3] Add a filter implementing HDR image reconstruction from a single exposure using deep CNNs

Submitted by Guo, Yejun on Oct. 19, 2018, 6:04 p.m.

Details

Message ID 1539972268-31472-1-git-send-email-yejun.guo@intel.com
State New
Headers show

Commit Message

Guo, Yejun Oct. 19, 2018, 6:04 p.m.
see the algorithm's paper and code below.

the filter's parameter looks like:
sdr2hdr=model_filename=/path_to_tensorflow_graph.pb:out_fmt=gbrp10le

The input of the deep CNN model is RGB24 while the output is float
for each color channel. This is the filter's default behavior to
output format with gbrpf32le. And gbrp10le is also supported as the
output, so we can see the rendering result in a player, as a reference.

To generate the model file, we need modify the original script a little.
- set name='y' for y_final within script at
https://github.com/gabrieleilertsen/hdrcnn/blob/master/network.py
- add the following code to the script at
https://github.com/gabrieleilertsen/hdrcnn/blob/master/hdrcnn_predict.py

graph = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, ["y"])
tf.train.write_graph(graph, '.', 'graph.pb', as_text=False)

The filter only works when tensorflow C api is supported in the system,
native backend is not supported since there are some different types of
layers in the deep CNN model, besides CONV and DEPTH_TO_SPACE.

https://arxiv.org/pdf/1710.07480.pdf:
  author       = "Eilertsen, Gabriel and Kronander, Joel, and Denes, Gyorgy and Mantiuk, Rafał and Unger, Jonas",
  title        = "HDR image reconstruction from a single exposure using deep CNNs",
  journal      = "ACM Transactions on Graphics (TOG)",
  number       = "6",
  volume       = "36",
  articleno    = "178",
  year         = "2017"

https://github.com/gabrieleilertsen/hdrcnn

btw, as a whole solution, metadata should also be generated from
the sdr video, so to be encoded as a HDR video. Not supported yet.
This patch just focuses on this paper.

v3: use int16_t instead of short
v2: use AV_OPT_TYPE_PIXEL_FMT for filter option
    remove some unnecessary code
    Use in->linesize[0] and FFMAX/FFMIN
    remove flag AVFILTER_FLAG_SLICE_THREADS
    add av_log message when error

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
---
 libavfilter/Makefile     |   1 +
 libavfilter/allfilters.c |   1 +
 libavfilter/vf_sdr2hdr.c | 266 +++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 268 insertions(+)
 create mode 100644 libavfilter/vf_sdr2hdr.c

Comments

Vittorio Giovara Oct. 19, 2018, 3:32 p.m.
On Fri, Oct 19, 2018 at 10:11 AM Guo, Yejun <yejun.guo@intel.com> wrote:

> see the algorithm's paper and code below.
>
> the filter's parameter looks like:
> sdr2hdr=model_filename=/path_to_tensorflow_graph.pb:out_fmt=gbrp10le
>

can you add some usage documentation to doc/filters.texi?

The input of the deep CNN model is RGB24 while the output is float
> for each color channel. This is the filter's default behavior to
> output format with gbrpf32le. And gbrp10le is also supported as the
> output, so we can see the rendering result in a player, as a reference.
>
> To generate the model file, we need modify the original script a little.
> - set name='y' for y_final within script at
> https://github.com/gabrieleilertsen/hdrcnn/blob/master/network.py
> - add the following code to the script at
> https://github.com/gabrieleilertsen/hdrcnn/blob/master/hdrcnn_predict.py
>
> graph = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def,
> ["y"])
> tf.train.write_graph(graph, '.', 'graph.pb', as_text=False)
>
> The filter only works when tensorflow C api is supported in the system,
> native backend is not supported since there are some different types of
> layers in the deep CNN model, besides CONV and DEPTH_TO_SPACE.
>
> https://arxiv.org/pdf/1710.07480.pdf:
>   author       = "Eilertsen, Gabriel and Kronander, Joel, and Denes,
> Gyorgy and Mantiuk, Rafał and Unger, Jonas",
>   title        = "HDR image reconstruction from a single exposure using
> deep CNNs",
>   journal      = "ACM Transactions on Graphics (TOG)",
>   number       = "6",
>   volume       = "36",
>   articleno    = "178",
>   year         = "2017"
>
> https://github.com/gabrieleilertsen/hdrcnn
>
> btw, as a whole solution, metadata should also be generated from
> the sdr video, so to be encoded as a HDR video. Not supported yet.
> This patch just focuses on this paper.
>

Is this something you are working on and will it be added later?


> v3: use int16_t instead of short
> v2: use AV_OPT_TYPE_PIXEL_FMT for filter option
>     remove some unnecessary code
>     Use in->linesize[0] and FFMAX/FFMIN
>     remove flag AVFILTER_FLAG_SLICE_THREADS
>     add av_log message when error
>

there is no need for this block to be left in the commit log


> Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
> ---
>  libavfilter/Makefile     |   1 +
>  libavfilter/allfilters.c |   1 +
>  libavfilter/vf_sdr2hdr.c | 266
> +++++++++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 268 insertions(+)
>  create mode 100644 libavfilter/vf_sdr2hdr.c
>
> +static av_cold int init(AVFilterContext* context)
> +{
> +    SDR2HDRContext* ctx = context->priv;
> +
> +    if (ctx->out_fmt != AV_PIX_FMT_GBRPF32LE && ctx->out_fmt !=
> AV_PIX_FMT_GBRP10LE) {
> +        av_log(context, AV_LOG_ERROR, "could not support the output
> format\n");
> +        return AVERROR(ENOSYS);
> +    }
> +
> +#if (CONFIG_LIBTENSORFLOW == 1)
> +    ctx->dnn_module = ff_get_dnn_module(DNN_TF);
> +    if (!ctx->dnn_module){
> +        av_log(context, AV_LOG_ERROR, "could not create DNN module for
> tensorflow backend\n");
> +        return AVERROR(ENOMEM);
> +    }
> +    if (!ctx->model_filename){
> +        av_log(context, AV_LOG_ERROR, "model file for network was not
> specified\n");
> +        return AVERROR(EIO);
> +    }
> +    if (!ctx->dnn_module->load_model) {
> +        av_log(context, AV_LOG_ERROR, "load_model for network was not
> specified\n");
> +        return AVERROR(EIO);
> +    }
> +    ctx->model = (ctx->dnn_module->load_model)(ctx->model_filename);
> +    if (!ctx->model){
> +        av_log(context, AV_LOG_ERROR, "could not load DNN model\n");
> +        return AVERROR(EIO);
> +    }
> +    return 0;
> +#else
> +    return AVERROR(EIO);
> +#endif
> +}
>

this is incorrect, what you should do is make libtensorflow a dependency of
this filter in the configure file and disable this filter when it is not
enabled


> +
> +static int query_formats(AVFilterContext* context)
> +{
> +    const enum AVPixelFormat in_formats[] = {AV_PIX_FMT_RGB24,
> +                                             AV_PIX_FMT_NONE};
> +    enum AVPixelFormat out_formats[2];
> +    SDR2HDRContext* ctx = context->priv;
> +    AVFilterFormats* formats_list;
> +    int ret = 0;
> +
> +    formats_list = ff_make_format_list(in_formats);
> +    if ((ret = ff_formats_ref(formats_list,
> &context->inputs[0]->out_formats)) < 0)
> +        return ret;
> +
> +    out_formats[0] = ctx->out_fmt;
> +    out_formats[1] = AV_PIX_FMT_NONE;
> +    formats_list = ff_make_format_list(out_formats);
> +    if ((ret = ff_formats_ref(formats_list,
> &context->outputs[0]->in_formats)) < 0)
> +        return ret;
> +
> +    return 0;
> +}
> +
> +static int config_props(AVFilterLink* inlink)
> +{
> +    AVFilterContext* context = inlink->dst;
> +    SDR2HDRContext* ctx = context->priv;
> +    AVFilterLink* outlink = context->outputs[0];
> +    DNNReturnType result;
> +
> +    // the dnn model is tied with resolution due to deconv layer of
> tensorflow
> +    // now just support 1920*1080 and so the magic numbers within this
> file
> +    if (inlink->w != 1920 || inlink->h != 1080) {
> +        av_log(context, AV_LOG_ERROR, "only support frame size with
> 1920*1080\n");
> +        return AVERROR(ENOSYS);
> +     }
>

is there any work planned to extend this to other resolutions?


> +
> +    ctx->input.width = 1920;
> +    ctx->input.height = 1088;  //the model requires height is a multiple
> of 32,
> +    ctx->input.channels = 3;
> +
> +    result = (ctx->model->set_input_output)(ctx->model->model,
> &ctx->input, &ctx->output);
> +    if (result != DNN_SUCCESS){
> +        av_log(context, AV_LOG_ERROR, "could not set input and output for
> the model\n");
> +        return AVERROR(EIO);
> +    }
> +
> +    memset(ctx->input.data, 0, ctx->input.channels * ctx->input.width *
> ctx->input.height * sizeof(float));
> +    outlink->h = 1080;
> +    outlink->w = 1920;
> +    return 0;
> +}
> +
> +static float qsort_comparison_function_float(const void *a, const void *b)
> +{
> +    return *(const float *)a - *(const float *)b;
> +}
> +
> +static int filter_frame(AVFilterLink* inlink, AVFrame* in)
> +{
> +    DNNReturnType dnn_result = DNN_SUCCESS;
> +    AVFilterContext* context = inlink->dst;
> +    SDR2HDRContext* ctx = context->priv;
> +    AVFilterLink* outlink = context->outputs[0];
> +    AVFrame* out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
> +    int total_pixels = in->height * in->width;
> +
> +    av_frame_copy_props(out, in);
>

check for allocation failures here


> +
> +    for (int i = 0; i < in->linesize[0] * in->height; ++i) {
> +        ctx->input.data[i] = in->data[0][i] / 255.0f;
> +    }
> +
> +    dnn_result = (ctx->dnn_module->execute_model)(ctx->model);
> +    if (dnn_result != DNN_SUCCESS){
> +        av_log(context, AV_LOG_ERROR, "failed to execute loaded model\n");
> +        return AVERROR(EIO);
> +    }
> +
> +    if (ctx->out_fmt == AV_PIX_FMT_GBRPF32LE) {
> +        float* outg = (float*)out->data[0];
> +        float* outb = (float*)out->data[1];
> +        float* outr = (float*)out->data[2];
> +        for (int i = 0; i < total_pixels; ++i) {
> +            float r = ctx->output.data[i*3];
> +            float g = ctx->output.data[i*3+1];
> +            float b = ctx->output.data[i*3+2];
> +            outr[i] = r;
> +            outg[i] = g;
> +            outb[i] = b;
> +        }
> +    } else {
> +        // here, we just use a rough mapping to the 10bit contents
> +        // meta data generation for HDR video encoding is not supported
> yet
> +        float* converted_data = (float*)malloc(total_pixels * 3 *
> sizeof(float));
>

don't use malloc, replace with av_malloc, same for free below


> +        int16_t* outg = (int16_t*)out->data[0];
> +        int16_t* outb = (int16_t*)out->data[1];
> +        int16_t* outr = (int16_t*)out->data[2];
> +
> +        float max = 1.0f;
> +        for (int i = 0; i < total_pixels * 3; ++i) {
> +            float d = ctx->output.data[i];
> +            d = sqrt(d);
> +            converted_data[i] = d;
> +            max = FFMAX(d, max);
> +        }
> +
> +        if (max > 1.0f) {
> +            AV_QSORT(converted_data, total_pixels * 3, float,
> qsort_comparison_function_float);
> +            // 0.5% pixels are clipped
> +            max = converted_data[(int)(total_pixels * 3 * 0.995)];
> +            max = FFMAX(max, 1.0f);
> +
> +            for (int i = 0; i < total_pixels * 3; ++i) {
> +                float d = ctx->output.data[i];
> +                d = sqrt(d);
> +                d = FFMIN(d, max);
> +                converted_data[i] = d;
> +            }
> +        }
> +
> +        for (int i = 0; i < total_pixels; ++i) {
> +            float r = converted_data[i*3];
> +            float g = converted_data[i*3+1];
> +            float b = converted_data[i*3+2];
> +            outr[i] = r / max * 1023;
> +            outg[i] = g / max * 1023;
> +            outb[i] = b / max * 1023;
> +        }
> +
> +        free(converted_data);
> +    }
> +
> +    av_frame_free(&in);
> +    return ff_filter_frame(outlink, out);
> +}
> +
> +static av_cold void uninit(AVFilterContext* context)
> +{
> +    SDR2HDRContext* ctx = context->priv;
> +
> +    if (ctx->dnn_module){
> +        (ctx->dnn_module->free_model)(&ctx->model);
> +        av_freep(&ctx->dnn_module);
> +    }
> +}
> +
> +static const AVFilterPad sdr2hdr_inputs[] = {
> +    {
> +        .name         = "default",
> +        .type         = AVMEDIA_TYPE_VIDEO,
> +        .config_props = config_props,
> +        .filter_frame = filter_frame,
> +    },
> +    { NULL }
> +};
> +
> +static const AVFilterPad sdr2hdr_outputs[] = {
> +    {
> +        .name = "default",
> +        .type = AVMEDIA_TYPE_VIDEO,
> +    },
> +    { NULL }
> +};
> +
> +AVFilter ff_vf_sdr2hdr = {
> +    .name          = "sdr2hdr",
> +    .description   = NULL_IF_CONFIG_SMALL("HDR image reconstruction from
> a single exposure using deep CNNs."),
>

why "reconstruction"? there is nothing to construct back if the source
wasn't hdr to begin with
"tonemap" is probably a better term here, in my opinion
same for previous uses
Guo, Yejun Oct. 22, 2018, 4:19 a.m.
From: Vittorio Giovara [mailto:vittorio.giovara@gmail.com]

Sent: Friday, October 19, 2018 11:33 PM
To: FFmpeg development discussions and patches <ffmpeg-devel@ffmpeg.org>
Cc: Guo, Yejun <yejun.guo@intel.com>; Guo@ffbox0-bg.ffmpeg.org
Subject: Re: [FFmpeg-devel] [PATCH V3] Add a filter implementing HDR image reconstruction from a single exposure using deep CNNs


On Fri, Oct 19, 2018 at 10:11 AM Guo, Yejun <yejun.guo@intel.com<mailto:yejun.guo@intel.com>> wrote:
see the algorithm's paper and code below.

the filter's parameter looks like:
sdr2hdr=model_filename=/path_to_tensorflow_graph.pb:out_fmt=gbrp10le

>  can you add some usage documentation to doc/filters.texi?


sure, will add it.

The input of the deep CNN model is RGB24 while the output is float
for each color channel. This is the filter's default behavior to
output format with gbrpf32le. And gbrp10le is also supported as the
output, so we can see the rendering result in a player, as a reference.

To generate the model file, we need modify the original script a little.
- set name='y' for y_final within script at
https://github.com/gabrieleilertsen/hdrcnn/blob/master/network.py
- add the following code to the script at
https://github.com/gabrieleilertsen/hdrcnn/blob/master/hdrcnn_predict.py

graph = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, ["y"])
tf.train.write_graph(graph, '.', 'graph.pb', as_text=False)

The filter only works when tensorflow C api is supported in the system,
native backend is not supported since there are some different types of
layers in the deep CNN model, besides CONV and DEPTH_TO_SPACE.

https://arxiv.org/pdf/1710.07480.pdf:
  author       = "Eilertsen, Gabriel and Kronander, Joel, and Denes, Gyorgy and Mantiuk, Rafał and Unger, Jonas",
  title        = "HDR image reconstruction from a single exposure using deep CNNs",
  journal      = "ACM Transactions on Graphics (TOG)",
  number       = "6",
  volume       = "36",
  articleno    = "178",
  year         = "2017"

https://github.com/gabrieleilertsen/hdrcnn

btw, as a whole solution, metadata should also be generated from
the sdr video, so to be encoded as a HDR video. Not supported yet.
This patch just focuses on this paper.

> Is this something you are working on and will it be added later?


yes, this is in our team’s todo list.

v3: use int16_t instead of short
v2: use AV_OPT_TYPE_PIXEL_FMT for filter option
    remove some unnecessary code
    Use in->linesize[0] and FFMAX/FFMIN
    remove flag AVFILTER_FLAG_SLICE_THREADS
    add av_log message when error

> there is no need for this block to be left in the commit log


ok, will remove it.

Signed-off-by: Guo, Yejun <yejun.guo@intel.com<mailto:yejun.guo@intel.com>>

---
 libavfilter/Makefile     |   1 +
 libavfilter/allfilters.c |   1 +
 libavfilter/vf_sdr2hdr.c | 266 +++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 268 insertions(+)
 create mode 100644 libavfilter/vf_sdr2hdr.c

+static av_cold int init(AVFilterContext* context)
+{
+    SDR2HDRContext* ctx = context->priv;
+
+    if (ctx->out_fmt != AV_PIX_FMT_GBRPF32LE && ctx->out_fmt != AV_PIX_FMT_GBRP10LE) {
+        av_log(context, AV_LOG_ERROR, "could not support the output format\n");
+        return AVERROR(ENOSYS);
+    }
+
+#if (CONFIG_LIBTENSORFLOW == 1)
+    ctx->dnn_module = ff_get_dnn_module(DNN_TF);
+    if (!ctx->dnn_module){
+        av_log(context, AV_LOG_ERROR, "could not create DNN module for tensorflow backend\n");
+        return AVERROR(ENOMEM);
+    }
+    if (!ctx->model_filename){
+        av_log(context, AV_LOG_ERROR, "model file for network was not specified\n");
+        return AVERROR(EIO);
+    }
+    if (!ctx->dnn_module->load_model) {
+        av_log(context, AV_LOG_ERROR, "load_model for network was not specified\n");
+        return AVERROR(EIO);
+    }
+    ctx->model = (ctx->dnn_module->load_model)(ctx->model_filename);
+    if (!ctx->model){
+        av_log(context, AV_LOG_ERROR, "could not load DNN model\n");
+        return AVERROR(EIO);
+    }
+    return 0;
+#else
+    return AVERROR(EIO);
+#endif
+}

> this is incorrect, what you should do is make libtensorflow a dependency of this filter in the configure file and disable this filter when it is not enabled


thanks, will fix it.

+
+static int query_formats(AVFilterContext* context)
+{
+    const enum AVPixelFormat in_formats[] = {AV_PIX_FMT_RGB24,
+                                             AV_PIX_FMT_NONE};
+    enum AVPixelFormat out_formats[2];
+    SDR2HDRContext* ctx = context->priv;
+    AVFilterFormats* formats_list;
+    int ret = 0;
+
+    formats_list = ff_make_format_list(in_formats);
+    if ((ret = ff_formats_ref(formats_list, &context->inputs[0]->out_formats)) < 0)
+        return ret;
+
+    out_formats[0] = ctx->out_fmt;
+    out_formats[1] = AV_PIX_FMT_NONE;
+    formats_list = ff_make_format_list(out_formats);
+    if ((ret = ff_formats_ref(formats_list, &context->outputs[0]->in_formats)) < 0)
+        return ret;
+
+    return 0;
+}
+
+static int config_props(AVFilterLink* inlink)
+{
+    AVFilterContext* context = inlink->dst;
+    SDR2HDRContext* ctx = context->priv;
+    AVFilterLink* outlink = context->outputs[0];
+    DNNReturnType result;
+
+    // the dnn model is tied with resolution due to deconv layer of tensorflow
+    // now just support 1920*1080 and so the magic numbers within this file
+    if (inlink->w != 1920 || inlink->h != 1080) {
+        av_log(context, AV_LOG_ERROR, "only support frame size with 1920*1080\n");
+        return AVERROR(ENOSYS);
+     }

> is there any work planned to extend this to other resolutions?


yes, the plan is to fix tensorflow first, to make deconv layer not tied with resolution. it is also in our team’s todo list.

+
+    ctx->input.width = 1920;
+    ctx->input.height = 1088;  //the model requires height is a multiple of 32,
+    ctx->input.channels = 3;
+
+    result = (ctx->model->set_input_output)(ctx->model->model, &ctx->input, &ctx->output);
+    if (result != DNN_SUCCESS){
+        av_log(context, AV_LOG_ERROR, "could not set input and output for the model\n");
+        return AVERROR(EIO);
+    }
+
+    memset(ctx->input.data, 0, ctx->input.channels * ctx->input.width * ctx->input.height * sizeof(float));
+    outlink->h = 1080;
+    outlink->w = 1920;
+    return 0;
+}
+
+static float qsort_comparison_function_float(const void *a, const void *b)
+{
+    return *(const float *)a - *(const float *)b;
+}
+
+static int filter_frame(AVFilterLink* inlink, AVFrame* in)
+{
+    DNNReturnType dnn_result = DNN_SUCCESS;
+    AVFilterContext* context = inlink->dst;
+    SDR2HDRContext* ctx = context->priv;
+    AVFilterLink* outlink = context->outputs[0];
+    AVFrame* out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
+    int total_pixels = in->height * in->width;
+
+    av_frame_copy_props(out, in);

> check for allocation failures here


thanks, will add the check.

+
+    for (int i = 0; i < in->linesize[0] * in->height; ++i) {
+        ctx->input.data[i] = in->data[0][i] / 255.0f;
+    }
+
+    dnn_result = (ctx->dnn_module->execute_model)(ctx->model);
+    if (dnn_result != DNN_SUCCESS){
+        av_log(context, AV_LOG_ERROR, "failed to execute loaded model\n");
+        return AVERROR(EIO);
+    }
+
+    if (ctx->out_fmt == AV_PIX_FMT_GBRPF32LE) {
+        float* outg = (float*)out->data[0];
+        float* outb = (float*)out->data[1];
+        float* outr = (float*)out->data[2];
+        for (int i = 0; i < total_pixels; ++i) {
+            float r = ctx->output.data[i*3];
+            float g = ctx->output.data[i*3+1];
+            float b = ctx->output.data[i*3+2];
+            outr[i] = r;
+            outg[i] = g;
+            outb[i] = b;
+        }
+    } else {
+        // here, we just use a rough mapping to the 10bit contents
+        // meta data generation for HDR video encoding is not supported yet
+        float* converted_data = (float*)malloc(total_pixels * 3 * sizeof(float));

> don't use malloc, replace with av_malloc, same for free below


thanks, will fix it.

+        int16_t* outg = (int16_t*)out->data[0];
+        int16_t* outb = (int16_t*)out->data[1];
+        int16_t* outr = (int16_t*)out->data[2];
+
+        float max = 1.0f;
+        for (int i = 0; i < total_pixels * 3; ++i) {
+            float d = ctx->output.data[i];
+            d = sqrt(d);
+            converted_data[i] = d;
+            max = FFMAX(d, max);
+        }
+
+        if (max > 1.0f) {
+            AV_QSORT(converted_data, total_pixels * 3, float, qsort_comparison_function_float);
+            // 0.5% pixels are clipped
+            max = converted_data[(int)(total_pixels * 3 * 0.995)];
+            max = FFMAX(max, 1.0f);
+
+            for (int i = 0; i < total_pixels * 3; ++i) {
+                float d = ctx->output.data[i];
+                d = sqrt(d);
+                d = FFMIN(d, max);
+                converted_data[i] = d;
+            }
+        }
+
+        for (int i = 0; i < total_pixels; ++i) {
+            float r = converted_data[i*3];
+            float g = converted_data[i*3+1];
+            float b = converted_data[i*3+2];
+            outr[i] = r / max * 1023;
+            outg[i] = g / max * 1023;
+            outb[i] = b / max * 1023;
+        }
+
+        free(converted_data);
+    }
+
+    av_frame_free(&in);
+    return ff_filter_frame(outlink, out);
+}
+
+static av_cold void uninit(AVFilterContext* context)
+{
+    SDR2HDRContext* ctx = context->priv;
+
+    if (ctx->dnn_module){
+        (ctx->dnn_module->free_model)(&ctx->model);
+        av_freep(&ctx->dnn_module);
+    }
+}
+
+static const AVFilterPad sdr2hdr_inputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_VIDEO,
+        .config_props = config_props,
+        .filter_frame = filter_frame,
+    },
+    { NULL }
+};
+
+static const AVFilterPad sdr2hdr_outputs[] = {
+    {
+        .name = "default",
+        .type = AVMEDIA_TYPE_VIDEO,
+    },
+    { NULL }
+};
+
+AVFilter ff_vf_sdr2hdr = {
+    .name          = "sdr2hdr",
+    .description   = NULL_IF_CONFIG_SMALL("HDR image reconstruction from a single exposure using deep CNNs."),

> why "reconstruction"? there is nothing to construct back if the source wasn't hdr to begin with

> "tonemap" is probably a better term here, in my opinion

> same for previous uses


there is more detail data generated with the dnn model, the model accepts sdr frame and generates hdr data.
see more detail in paper @https://arxiv.org/pdf/1710.07480.pdf, and the description comes from the title of this paper.


--
Vittorio
Vittorio Giovara Oct. 22, 2018, 4:30 a.m.
On Mon, Oct 22, 2018 at 4:19 AM Guo, Yejun <yejun.guo@intel.com> wrote:

> +    .description   = NULL_IF_CONFIG_SMALL("HDR image reconstruction from
> a single exposure using deep CNNs."),
>
>
>
> > why "reconstruction"? there is nothing to construct back if the source
> wasn't hdr to begin with
>
> > "tonemap" is probably a better term here, in my opinion
>
> > same for previous uses
>
>
>
> there is more detail data generated with the dnn model, the model accepts
> sdr frame and generates hdr data.
>
> see more detail in paper @https://arxiv.org/pdf/1710.07480.pdf, and the
> description comes from the title of this paper.
>

Thanks for the link, however i'm still not sold on the term. You "generate"
hdr data, not "reconstruct": it's generated/estimated/made up data, not
data that is lost and needs to be reconstrcuted. I suggested "tonemap"
because you're mapping SDR tones (aka colors) to HDR ones, and that seems
the right term to use. If you really dislike it, at least consider "HDR
image generation from a single exposure using deep CNNs" which would work
much better.
Guo, Yejun Oct. 22, 2018, 8:56 a.m.
From: Vittorio Giovara [mailto:vittorio.giovara@gmail.com]

Sent: Monday, October 22, 2018 12:30 PM
To: Guo, Yejun <yejun.guo@intel.com>
Cc: FFmpeg development discussions and patches <ffmpeg-devel@ffmpeg.org>
Subject: Re: [FFmpeg-devel] [PATCH V3] Add a filter implementing HDR image reconstruction from a single exposure using deep CNNs


On Mon, Oct 22, 2018 at 4:19 AM Guo, Yejun <yejun.guo@intel.com<mailto:yejun.guo@intel.com>> wrote:
+    .description   = NULL_IF_CONFIG_SMALL("HDR image reconstruction from a single exposure using deep CNNs."),

> why "reconstruction"? there is nothing to construct back if the source wasn't hdr to begin with

> "tonemap" is probably a better term here, in my opinion

> same for previous uses


there is more detail data generated with the dnn model, the model accepts sdr frame and generates hdr data.
see more detail in paper @https://arxiv.org/pdf/1710.07480.pdf, and the description comes from the title of this paper.

> Thanks for the link, however i'm still not sold on the term. You "generate" hdr data, not "reconstruct": it's generated/estimated/made up data, not data that is lost and needs to be reconstrcuted. I suggested "tonemap" because you're mapping SDR tones (aka colors) to HDR ones, and that seems the right term to use. If you really dislike it, at least consider "HDR image generation from a single exposure using deep CNNs" which would work much better.


got your point, it’s nice, thanks, will choose ‘generation’

--
Vittorio
Ruiling Song Oct. 22, 2018, 2:02 p.m.
> Thanks for the link, however i'm still not sold on the term. You "generate"

> hdr data, not "reconstruct": it's generated/estimated/made up data, not

> data that is lost and needs to be reconstrcuted. I suggested "tonemap"

> because you're mapping SDR tones (aka colors) to HDR ones, and that seems

> the right term to use. If you really dislike it, at least consider "HDR

> image generation from a single exposure using deep CNNs" which would work

I think "inverse/reverse tone mapping" looks a little better. I see many papers use this term when talking about sdr to hdr.

> much better.

> --

> Vittorio

> _______________________________________________

> ffmpeg-devel mailing list

> ffmpeg-devel@ffmpeg.org

> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Guo, Yejun Oct. 22, 2018, 2:36 p.m.
> -----Original Message-----

> From: Song, Ruiling

> Sent: Monday, October 22, 2018 10:02 PM

> To: FFmpeg development discussions and patches <ffmpeg-

> devel@ffmpeg.org>; Guo, Yejun <yejun.guo@intel.com>

> Subject: RE: [FFmpeg-devel] [PATCH V3] Add a filter implementing HDR

> image reconstruction from a single exposure using deep CNNs

> 

> > Thanks for the link, however i'm still not sold on the term. You "generate"

> > hdr data, not "reconstruct": it's generated/estimated/made up data,

> > not data that is lost and needs to be reconstrcuted. I suggested "tonemap"

> > because you're mapping SDR tones (aka colors) to HDR ones, and that

> > seems the right term to use. If you really dislike it, at least

> > consider "HDR image generation from a single exposure using deep CNNs"

> > which would work

> I think "inverse/reverse tone mapping" looks a little better. I see many

> papers use this term when talking about sdr to hdr.

> 


I personally more like "HDR image generation from a single exposure using deep CNNs" to 
bind it loosely with the title of the paper.

> > much better.

> > --

> > Vittorio

> > _______________________________________________

> > ffmpeg-devel mailing list

> > ffmpeg-devel@ffmpeg.org

> > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Patch hide | download patch | download mbox

diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 62cc2f5..88e7da6 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -360,6 +360,7 @@  OBJS-$(CONFIG_SOBEL_OPENCL_FILTER)           += vf_convolution_opencl.o opencl.o
 OBJS-$(CONFIG_SPLIT_FILTER)                  += split.o
 OBJS-$(CONFIG_SPP_FILTER)                    += vf_spp.o
 OBJS-$(CONFIG_SR_FILTER)                     += vf_sr.o
+OBJS-$(CONFIG_SDR2HDR_FILTER)                += vf_sdr2hdr.o
 OBJS-$(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 5e72803..1645c0f 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -319,6 +319,7 @@  extern AVFilter ff_vf_scale_npp;
 extern AVFilter ff_vf_scale_qsv;
 extern AVFilter ff_vf_scale_vaapi;
 extern AVFilter ff_vf_scale2ref;
+extern AVFilter ff_vf_sdr2hdr;
 extern AVFilter ff_vf_select;
 extern AVFilter ff_vf_selectivecolor;
 extern AVFilter ff_vf_sendcmd;
diff --git a/libavfilter/vf_sdr2hdr.c b/libavfilter/vf_sdr2hdr.c
new file mode 100644
index 0000000..6a51a54
--- /dev/null
+++ b/libavfilter/vf_sdr2hdr.c
@@ -0,0 +1,266 @@ 
+/*
+ * Copyright (c) 2018 Guo Yejun
+ *
+ * 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
+ * Filter implementing HDR image reconstruction from a single exposure using deep CNNs.
+ * https://arxiv.org/pdf/1710.07480.pdf
+ */
+
+#include "avfilter.h"
+#include "formats.h"
+#include "internal.h"
+#include "libavutil/opt.h"
+#include "libavutil/qsort.h"
+#include "libavformat/avio.h"
+#include "libswscale/swscale.h"
+#include "dnn_interface.h"
+#include <math.h>
+
+typedef struct SDR2HDRContext {
+    const AVClass *class;
+
+    char* model_filename;
+    enum AVPixelFormat out_fmt;
+    DNNModule* dnn_module;
+    DNNModel* model;
+    DNNData input, output;
+} SDR2HDRContext;
+
+#define OFFSET(x) offsetof(SDR2HDRContext, x)
+#define FLAGS AV_OPT_FLAG_FILTERING_PARAM | AV_OPT_FLAG_VIDEO_PARAM
+static const AVOption sdr2hdr_options[] = {
+    { "model_filename", "path to model file specifying network architecture and its parameters", OFFSET(model_filename), AV_OPT_TYPE_STRING, {.str=NULL}, 0, 0, FLAGS },
+    { "out_fmt", "the data format of the filter's output, it could be gbrpf32le [default] or gbrp10le", OFFSET(out_fmt), AV_OPT_TYPE_PIXEL_FMT, {.i64=AV_PIX_FMT_GBRPF32LE}, AV_PIX_FMT_NONE, AV_PIX_FMT_NB, FLAGS },
+    { NULL }
+};
+
+AVFILTER_DEFINE_CLASS(sdr2hdr);
+
+static av_cold int init(AVFilterContext* context)
+{
+    SDR2HDRContext* ctx = context->priv;
+
+    if (ctx->out_fmt != AV_PIX_FMT_GBRPF32LE && ctx->out_fmt != AV_PIX_FMT_GBRP10LE) {
+        av_log(context, AV_LOG_ERROR, "could not support the output format\n");
+        return AVERROR(ENOSYS);
+    }
+
+#if (CONFIG_LIBTENSORFLOW == 1)
+    ctx->dnn_module = ff_get_dnn_module(DNN_TF);
+    if (!ctx->dnn_module){
+        av_log(context, AV_LOG_ERROR, "could not create DNN module for tensorflow backend\n");
+        return AVERROR(ENOMEM);
+    }
+    if (!ctx->model_filename){
+        av_log(context, AV_LOG_ERROR, "model file for network was not specified\n");
+        return AVERROR(EIO);
+    }
+    if (!ctx->dnn_module->load_model) {
+        av_log(context, AV_LOG_ERROR, "load_model for network was not specified\n");
+        return AVERROR(EIO);
+    }
+    ctx->model = (ctx->dnn_module->load_model)(ctx->model_filename);
+    if (!ctx->model){
+        av_log(context, AV_LOG_ERROR, "could not load DNN model\n");
+        return AVERROR(EIO);
+    }
+    return 0;
+#else
+    return AVERROR(EIO);
+#endif
+}
+
+static int query_formats(AVFilterContext* context)
+{
+    const enum AVPixelFormat in_formats[] = {AV_PIX_FMT_RGB24,
+                                             AV_PIX_FMT_NONE};
+    enum AVPixelFormat out_formats[2];
+    SDR2HDRContext* ctx = context->priv;
+    AVFilterFormats* formats_list;
+    int ret = 0;
+
+    formats_list = ff_make_format_list(in_formats);
+    if ((ret = ff_formats_ref(formats_list, &context->inputs[0]->out_formats)) < 0)
+        return ret;
+
+    out_formats[0] = ctx->out_fmt;
+    out_formats[1] = AV_PIX_FMT_NONE;
+    formats_list = ff_make_format_list(out_formats);
+    if ((ret = ff_formats_ref(formats_list, &context->outputs[0]->in_formats)) < 0)
+        return ret;
+
+    return 0;
+}
+
+static int config_props(AVFilterLink* inlink)
+{
+    AVFilterContext* context = inlink->dst;
+    SDR2HDRContext* ctx = context->priv;
+    AVFilterLink* outlink = context->outputs[0];
+    DNNReturnType result;
+
+    // the dnn model is tied with resolution due to deconv layer of tensorflow
+    // now just support 1920*1080 and so the magic numbers within this file
+    if (inlink->w != 1920 || inlink->h != 1080) {
+        av_log(context, AV_LOG_ERROR, "only support frame size with 1920*1080\n");
+        return AVERROR(ENOSYS);
+     }
+
+    ctx->input.width = 1920;
+    ctx->input.height = 1088;  //the model requires height is a multiple of 32,
+    ctx->input.channels = 3;
+
+    result = (ctx->model->set_input_output)(ctx->model->model, &ctx->input, &ctx->output);
+    if (result != DNN_SUCCESS){
+        av_log(context, AV_LOG_ERROR, "could not set input and output for the model\n");
+        return AVERROR(EIO);
+    }
+
+    memset(ctx->input.data, 0, ctx->input.channels * ctx->input.width * ctx->input.height * sizeof(float));
+    outlink->h = 1080;
+    outlink->w = 1920;
+    return 0;
+}
+
+static float qsort_comparison_function_float(const void *a, const void *b)
+{
+    return *(const float *)a - *(const float *)b;
+}
+
+static int filter_frame(AVFilterLink* inlink, AVFrame* in)
+{
+    DNNReturnType dnn_result = DNN_SUCCESS;
+    AVFilterContext* context = inlink->dst;
+    SDR2HDRContext* ctx = context->priv;
+    AVFilterLink* outlink = context->outputs[0];
+    AVFrame* out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
+    int total_pixels = in->height * in->width;
+
+    av_frame_copy_props(out, in);
+
+    for (int i = 0; i < in->linesize[0] * in->height; ++i) {
+        ctx->input.data[i] = in->data[0][i] / 255.0f;
+    }
+
+    dnn_result = (ctx->dnn_module->execute_model)(ctx->model);
+    if (dnn_result != DNN_SUCCESS){
+        av_log(context, AV_LOG_ERROR, "failed to execute loaded model\n");
+        return AVERROR(EIO);
+    }
+
+    if (ctx->out_fmt == AV_PIX_FMT_GBRPF32LE) {
+        float* outg = (float*)out->data[0];
+        float* outb = (float*)out->data[1];
+        float* outr = (float*)out->data[2];
+        for (int i = 0; i < total_pixels; ++i) {
+            float r = ctx->output.data[i*3];
+            float g = ctx->output.data[i*3+1];
+            float b = ctx->output.data[i*3+2];
+            outr[i] = r;
+            outg[i] = g;
+            outb[i] = b;
+        }
+    } else {
+        // here, we just use a rough mapping to the 10bit contents
+        // meta data generation for HDR video encoding is not supported yet
+        float* converted_data = (float*)malloc(total_pixels * 3 * sizeof(float));
+        int16_t* outg = (int16_t*)out->data[0];
+        int16_t* outb = (int16_t*)out->data[1];
+        int16_t* outr = (int16_t*)out->data[2];
+
+        float max = 1.0f;
+        for (int i = 0; i < total_pixels * 3; ++i) {
+            float d = ctx->output.data[i];
+            d = sqrt(d);
+            converted_data[i] = d;
+            max = FFMAX(d, max);
+        }
+
+        if (max > 1.0f) {
+            AV_QSORT(converted_data, total_pixels * 3, float, qsort_comparison_function_float);
+            // 0.5% pixels are clipped
+            max = converted_data[(int)(total_pixels * 3 * 0.995)];
+            max = FFMAX(max, 1.0f);
+
+            for (int i = 0; i < total_pixels * 3; ++i) {
+                float d = ctx->output.data[i];
+                d = sqrt(d);
+                d = FFMIN(d, max);
+                converted_data[i] = d;
+            }
+        }
+
+        for (int i = 0; i < total_pixels; ++i) {
+            float r = converted_data[i*3];
+            float g = converted_data[i*3+1];
+            float b = converted_data[i*3+2];
+            outr[i] = r / max * 1023;
+            outg[i] = g / max * 1023;
+            outb[i] = b / max * 1023;
+        }
+
+        free(converted_data);
+    }
+
+    av_frame_free(&in);
+    return ff_filter_frame(outlink, out);
+}
+
+static av_cold void uninit(AVFilterContext* context)
+{
+    SDR2HDRContext* ctx = context->priv;
+
+    if (ctx->dnn_module){
+        (ctx->dnn_module->free_model)(&ctx->model);
+        av_freep(&ctx->dnn_module);
+    }
+}
+
+static const AVFilterPad sdr2hdr_inputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_VIDEO,
+        .config_props = config_props,
+        .filter_frame = filter_frame,
+    },
+    { NULL }
+};
+
+static const AVFilterPad sdr2hdr_outputs[] = {
+    {
+        .name = "default",
+        .type = AVMEDIA_TYPE_VIDEO,
+    },
+    { NULL }
+};
+
+AVFilter ff_vf_sdr2hdr = {
+    .name          = "sdr2hdr",
+    .description   = NULL_IF_CONFIG_SMALL("HDR image reconstruction from a single exposure using deep CNNs."),
+    .priv_size     = sizeof(SDR2HDRContext),
+    .init          = init,
+    .uninit        = uninit,
+    .query_formats = query_formats,
+    .inputs        = sdr2hdr_inputs,
+    .outputs       = sdr2hdr_outputs,
+    .priv_class    = &sdr2hdr_class,
+    .flags         = AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC,
+};