Multi-Object tracking using all variants of deepsort

Installation

• Install StrongSORT

  cd StrongSORT
  pip install -r requirements.txt

• Install ByteTrack

 cd ByteTrack
 pip install -r requirements.txt
 python3 setup.py develop

• Install SAHI

  git clone https://github.com/kadirnar/Yolov7-SAHI.git
  python3 setup.py install

Download Models

• Download Yolov7-E6E and put it under StrongSORT/weigths and ByteTrack/models folders respectively
• Download bytetrack_x_mot20 and put it under pretrained folder

Inference

DeepSort

cd StrongSORT
python deep_track_v7.py --yolo-weights weights/yolov7-e6e.pt --source train_part7/07_University_Campus --img-size 5120 --classes 0 --save-txt --save-img

StronSORT

cd StrongSORT
python track_v7.py --yolo-weights weights/yolov7-e6e.pt --source train_part7/07_University_Campus --img-size 5120 --classes 0 --save-txt --save-img

ByteTrack

cd ByteTrack
python tools/demo_track.py -f exps/example/mot/yolox_x_mix_mot20_ch.py -c pretrained/bytetrack_x_mot20.tar --fuse --save_result --path train_part7/07_University_Campus --conf 0.25 --nms 0.45 --tsize 5120

SAHI with ByteTrack (Yolov7)

cd ByteTrack
python tools/sahi_track_v7.py -f exps/example/mot/yolox_x_mix_mot20_ch.py -c pretrained/bytetrack_x_mot20.tar --path train_part7/07_University_Campus --tsize 5120 --save_result

SAHI with ByteTrack (YolovX)

Edit line 117-118 of tools/sahi_track_vX.py for changing sahi overlap threshold

cd ByteTrack
python tools/sahi_track_vX.py -f exps/example/mot/yolox_x_mix_mot20_ch.py -c pretrained/bytetrack_x_mot20.tar --path train_part7/07_University_Campus --tsize 5120 --conf 0.10 --nms 0.1 --save_result