Skip to content

Latest commit

 

History

History
32 lines (21 loc) · 852 Bytes

README.md

File metadata and controls

32 lines (21 loc) · 852 Bytes

gbssplime

Generalized Background Subtraction using Super-pixels and Label Integrated Motion Estimation

Jongwoo Lim, Bohyung Han, “Generalized Background Subtraction using Superpixels with Label Integrated Motion Estimation,” in ECCV 2014

Project webpage: http://cvlab.hanyang.ac.kr/proj/gbssplime/

How to run the code

  1. Build the two mex files - Coarse2FineTwoFrames.mex and mex_ersv.mex.

  2. Run 'Run.m' and choose the dataset to process.

  • The test video data will be automatically downloaded.

Parameters

SetupSequence.m:

  • vdir = './data/'; : where the dataset is stored.
  • rpath = ['../Result/', num2str(vid), '/']; : where the result is stored.

Run.m

  • SAVE_PROGRESS_IMAGES = true; : whether to save intermediate results.

SetupUnifiedParam.m

  • a whole bunch of parameters for the algorithm.