vtddggg / Adapted_TIRtracking

Adapting deep visible trackers for thermal infrared tracking

Adapted_TIRtracking

This demo is only used for test our algorithm for adapting deep visible trackers to thermal infrared domain. Here we adopt HCF as the base tracker.

Usage

1.Clone the repository

$ git clone https://github.com/xfmao/Adapted_TIRtracking.git

2.Download MatConvNet and extract it under the directory .\external\matconvnet.

Then precompile it by

> cd <MatConvNet>
> addpath matlab
> vl_compilenn

If you want to run the code with GPU, make sure that you have installed CUDA and cuDNN. And compile MatConvNet with cuDNN support.(refer to http://www.vlfeat.org/matconvnet/install/)

3.Download the VGG-Net-19 model and put it in .\model\.

4.Run demo.m.

Results

HCF without finetuning:

HCF after finetuning:

Citation

@inproceedings{Ma-ICCV-2015,
    title={Hierarchical Convolutional Features for Visual Tracking},
    Author = {Ma, Chao and Huang, Jia-Bin and Yang, Xiaokang and Yang, Ming-Hsuan},
    booktitle = {Proceedings of the IEEE International Conference on Computer Vision},
    pages={},
    Year = {2015}
}