TensorFlow port of the tracking method described in the paper Fully-Convolutional Siamese nets for object tracking.

This repo adds multiple object tracking, tracking at different scales, optimized inference time, easy API as well as configurable options.

Note1: at the moment this code only allows to use a pretrained net in forward mode.

See requirements.txt

Usage: python run_tracker_demo.py -cfg [config_file] -f [folder of video images]

Example: python run_tracker_demo.py -cfg ./config/design_ori.json -f ./data/interoll2

Make sure 'net_path' in the config file points to the path of your pretrained net.

Add training code

This repo is a full reimplementation of Fully-Convolutional Siamese Networks for Object Tracking (see paper https://www.robots.ox.ac.uk/~luca/siamese-fc.html)

Based off the original repository (https://github.com/torrvision/siamfc-tf)