ybsong00 / CREST-Release

CREST: Convolutional Residual Learning for Visual Tracking

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

This is the implementation of our CREST paper. The project page can be found here: https://ybsong00.github.io/iccv17/index.html

There are three folders in this repository where the matconvnet toolbox and Skiing sequences are contained. Our main development is kept in the folder CREST.

Before running our code, check if you have a state-of-the-art GPU. I develop this code using Titan Black. Make sure yours are better than mine :-).

Please download the VGG-16 model and put it under 'CREST/exp/model/'. You can download VGG-16 model via http://www.vlfeat.org/matconvnet/pretrained/.

Meanwhile, please configure matconvnet on your side. (You need to compile matconvnet using the provided because of the modifications.)

Try 'CREST/demo.m' to see the tracker performance on the Skiing sequences.

If you find the code useful, please cite:

    author    = {Song, Yibing and Ma, Chao and Gong, Lijun and Zhang, Jiawei and Lau, Rynson and Yang, Ming-Hsuan}, 
    title     = {CREST: Convolutional Residual Learning for Visual Tracking}, 
    booktitle = {IEEE International Conference on Computer Vision},
    pages     = {2555 -- 2564},
    year      = {2017}

You might encounter the vl_nnconv problem when using the code in windows. This code is developed in linux and the compatible issues exist between different OS. Check the following link for a solution.



CREST: Convolutional Residual Learning for Visual Tracking


Language:MATLAB 34.0%Language:Cuda 31.3%Language:C++ 13.8%Language:TeX 9.7%Language:Python 6.0%Language:Shell 1.9%Language:C 1.3%Language:Makefile 1.1%Language:HTML 0.5%Language:CSS 0.4%Language:JavaScript 0.0%Language:M 0.0%