This is dense_flow(yjxiong) project for opencv-3.1. https://github.com/yjxiong/dense_flow
Extracting dense flow field given a video.
- LibZip:
to install on ubuntu
apt-get install libzip-dev
on macbrew install libzip
git clone --recursive http://github.com/yjxiong/dense_flow
mkdir build && cd build
cmake .. && make -j
./extract_gpu -f test.avi -x tmp/flow_x -y tmp/flow_y -i tmp/image -b 20 -t 1 -d 0 -s 1 -o dir
test.avi
: input videotmp
: folder containing RGB images and optical flow imagesdir
: output generated images to folder. if set tozip
, will write images to zip files instead.
The warp optical flow is used in the following paper
@inproceedings{TSN2016ECCV,
author = {Limin Wang and
Yuanjun Xiong and
Zhe Wang and
Yu Qiao and
Dahua Lin and
Xiaoou Tang and
Luc {Van Gool}},
title = {Temporal Segment Networks: Towards Good Practices for Deep Action Recognition},