Supplementary Material for the ACCV 2020 Paper Entitled:
Video-Based Crowd Counting Using a Multi-Scale Optical Flow Pyramid Network
Mohammad Asiful Hossain, Kevin Cannons, Daesik Jang, Fabio Cuzzolin, and Zhan Xu

The supplementary material consists of two files:
1) This README.txt
2) A video file entitled Supplementary_Material.mp4

The supplementary video shows additional qualitative results for crowd density estimation in images and videos using the proposed methods from the paper. It was encoded with the H.264 codec and uses an MPEG-4 part 14 container. 

Note that in the video:
GT = ground truth
Baseline = Proposed baseline (w/o optical flow)
Proposed = Full proposed model (MOPN)
