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.
Python2.7 C++14 Matlab 2016a+
-
Clone this repository with
git clone https://github.com/gbc8181/TISLF.git
. -
Install Parallel Computing Toolbox in Matlab.
-
After toolbox is set up, you need to download a testing video (about 78M) from Youtube, and testing target images from Google Drive
-
Now, enter
oak/
directory of the codes. Runpython divide.py
to generate source images from video. -
Enter
pine/
directory of the codes. Runpartdivide.m
to segment video into scenes. -
Move
Num.mat
frompine/
towattle/
and runmain.m
Be sure you are in the root directory of the codes. Just try to run run.sh
The codes are released under the MIT License.