Final Project - Task 1
Pengwei Song & Yifan Qi
Model & Data
We use a mask R-CNN trained on Cityscapes from mmdetection to do inference on video. The model and config file can be downloaded from mmdet.
The test video can be downloaded from video.mp4.
Test
python test.py --file video.mp4 --out new_video.mp4