Skip to content
Code of MKCT-Tracker v1.0 (Matlab Version for Discussion)
MATLAB Objective-C C++ M
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
UAV123_10fps
external_libs
implementation
model
result
utils
.gitignore
MKCT_Demo.m
README.md
setup_paths.m

README.md

MKCT-tracker

Test passed
matlab-2017b MatConvNet-1.0--beta25 CUDA-8.0

Matlab implementation of Robust multi-kernelized correlators for UAV tracking with adaptive context analysis and dynamic weighted filters (MKCT-tracker).

Results

The following are the results from the experiment conducted on 100 challenging sequences extracted from UAV123@10fps.

Prec

Succ

Instructions

  1. Enter /model folder and run vgg_19_dl.m to download imagenet-vgg-verydeep-19.mat from here.
  2. Download matconvnet toolbox here and put it in /external.
  3. Run MKCT_Demo.m

Note: the default demo is using CPU to run the whole program. If GPU is required, just change false in the following lines in MKCT_Demo.m to true:

vl_compilenn('enableGpu',true,... 
             'cudaRoot', '<your-cuda-root>' ...
            );
vl_setupnn();

global enableGPU;
enableGPU = false;

Acknowledgements

Partly borrowed from KCC.

You can’t perform that action at this time.