Skip to content

Latest commit

 

History

History
46 lines (35 loc) · 1007 Bytes

README.md

File metadata and controls

46 lines (35 loc) · 1007 Bytes

ObjectTracking_CNN

A KCF tracker implementation which uses cnn feature as core.

Inspired

This project was inspired from tiago-cerveira/KCF-Tracker-CNN-Features.But change the backend from theano + lasagne to pytorch.And add muti-target tracing function.

Details

  • Support Muti-Target tracking
  • Support tripwire and restricted zone checking
  • Support CUDA accleration
  • Use PyTorch as cnn backend

Requirement

Runtime

  • Anaconda: Python 3.7.5

Library Requirement

From Python pip

  • numpy == any
  • opencv == 3.4.5.20

From Anaconda

  • PyTorch == any

CUDA (Optional)

  • CUDA == 9.1
  • cuDNN == 9.1

Useage

Run main

python main.py <path of the video file>

Use KCFMultiTracker

tracker = KCFMultiTracker(True, True)
tracker.init(rois, frist_frame)
while not_ended:
    success, result_rois = tracker.update(next_frame)
    next_frame = get_next_frame()