Target Image Video Search Based on Local Features (TISLF)

The codes are used for implementing TISLF for target image deterction in videos in:

B. Guan, H. Ye, H. Liu and W. A. Sethares, "Video Logo Retrieval Based on Local Features," 2020 IEEE International Conference on Image Processing (ICIP), Abu Dhabi, United Arab Emirates, 2020, pp. 1396-1400, doi: 10.1109/ICIP40778.2020.9191208.

Required softwares

Python2.7 C++14 Matlab 2016a+

How to run this code

1. Clone this repository with git clone https://github.com/gbc8181/TISLF.git.

2. Install Parallel Computing Toolbox in Matlab.

3. After toolbox is set up, you need to download a testing video (about 78M) from Youtube, and testing target images from Google Drive

4. Now, enter oak/ directory of the codes. Run python divide.py to generate source images from video.

5. Enter pine/ directory of the codes. Run partdivide.m to segment video into scenes.

6. Move Num.mat from pine/ to wattle/ and run main.m

How to run quickly

Be sure you are in the root directory of the codes. Just try to run run.sh

License

The codes are released under the MIT License.