[FFmpeg-devel] [PATCH V2] Add a filter implementing HDR image reconstruction from a single exposure using deep CNNs
Guo, Yejun
yejun.guo at intel.com
Thu Oct 18 18:09:37 EEST 2018
Hi,
Besides conv layer, this model uses these operations/layers:
PoolLayer, split, ConcatLayer, BatchNormLayer, DeConv2dLayer, reduce_max,
maximum, mul, pow, log, exp, relu, add, minimum, reshape, tile,
see detail in https://github.com/gabrieleilertsen/hdrcnn/blob/master/network.py
> -----Original Message-----
> From: ffmpeg-devel [mailto:ffmpeg-devel-bounces at ffmpeg.org] On Behalf
> Of Pedro Arthur
> Sent: Thursday, October 18, 2018 2:15 AM
> To: FFmpeg development discussions and patches <ffmpeg-
> devel at ffmpeg.org>
> Subject: Re: [FFmpeg-devel] [PATCH V2] Add a filter implementing HDR
> image reconstruction from a single exposure using deep CNNs
>
> Hi,
>
> How hard is it to support the native backend? which operations are missing
> or any other limitations?
>
> Em qua, 17 de out de 2018 às 05:47, Guo, Yejun <yejun.guo at intel.com>
> escreveu:
> >
> > see the algorithm's paper and code below.
> >
> > the filter's parameter looks like:
> >
> sdr2hdr=model_filename=/path_to_tensorflow_graph.pb:out_fmt=gbrp10l
> e
> >
> > 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.
> >
> > 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 at 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
> >
> > 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..fa61bfa
> > --- /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 },
> I think you could use "model" instead of "model_filename", it is shorter and
> more consistent with the vf_sr filter.
>
> > + { "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));
> > + short* outg = (short*)out->data[0];
> > + short* outb = (short*)out->data[1];
> > + short* outr = (short*)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,
> > +};
> > --
> > 2.7.4
> >
> > _______________________________________________
> > ffmpeg-devel mailing list
> > ffmpeg-devel at ffmpeg.org
> > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
More information about the ffmpeg-devel
mailing list