ziliharvey / lucas-kanade-tracking-and-correlation-filters Goto Github PK

View Code? Open in Web Editor NEW

47.0 1.0 14.0 93.77 MB

An implementation of several tracking algorithms based on Lucas Kanade algorithms

License: MIT License

Python 100.00%

computer-vision tracking-algorithm lucas-kanade optical-flow image-alignment

lucas-kanade-tracking-and-correlation-filters's People

Contributors

Stargazers

Watchers

Forkers

lingguomeng palelu81 yichunk weijiagao uhq1 yunajiang capri2014 caosen469 laokpa praguna moqimubai manishpaliwal1 tianvv1234

lucas-kanade-tracking-and-correlation-filters's Issues

Shouldn't it be (W[1,0] * x2 + W[1,1] * y2 + W[1,2])?

Lucas-Kanade-tracking-and-Correlation-Filters/src/LucasKanadeAffine.py

Line 27 in 9caca36

y2_w = W[0,0] * x2 + W[0,1] * y2 + W[0,2]

what should i change if i want to use user defined ROI (like the rect in LucasKanade) in the affine lucas-kanade

LucasKanadeAffine threshold bug

Hi! First, thank you for sharing your code. I tried to run your "testAerialSequence" code on my machine but found some strange point.
In LucasKanadeAffine.py, threshold is initialized with 5 and it always makes following while loop condition as False.
So LucasKanadeAffine function always returns [[1, 0, 0], [0, 1, 0]] and it means just identity matrix.
Can you please check this part and let me know how to fix this part to run your code.
Thanks!

Recommend Projects