HonglinChu / NanoTrack

Deep learning-based mobile model deployment(Object Tracking). Lightweight Object Tracking, NCNN,

NanoTrack

• NanoTrack is a lightweight and high speed tracking network which mainly referring to SiamBAN and LightTrack. It is suitable for deployment on embedded or mobile devices. In fact, it can run at >120FPS on Apple M1 CPU.

• Experiments show that NanoTrack algorithm has good performance on tracking datasets.

  Trackers	Backbone	ModeSize	VOT2018 EAO	VOT2019 EAO	GOT-10k-Val AO	GOT-10k-Val SR	DTB70 Success	DTB70 Precision
  NanoTrack	MobileNetV3	2.2MB	0.311	0.247	0.604	0.724	0.532	0.727
  CVPR2021 LightTrack	MobileNetV3	7.7MB	0.418	0.328	0.75	0.877	0.591	0.766
  WACV2022 SiamTPN	ShuffleNetV2	62.2MB	0.191	0.209	0.728	0.865	0.572	0.728
  ICRA2021 SiamAPN	AlexNet	118.7MB	0.248	0.235	0.622	0.708	0.585	0.786
  IROS2021 SiamAPN++	AlexNet	187MB	0.268	0.234	0.635	0.73	0.594	0.791

• We provide Android demo and MacOS demo based on ncnn inference framework.

• We also provide PyTorch code. It is friendly for training with much lower GPU memory cost than other models. NanoTrack only uses GOT-10k dataset to train, which only takes two hours on GPU3090.

Mac

PC demo

• 1. Modify your own CMakeList.txt
• 2. Build (Apple M1 CPU)

  $ sh make_macos_arm64.sh 
  

Android

Android demo

• 1. Modify your own CMakeList.txt
• 2. Download(password: 6cdd) OpenCV and NCNN libraries for Android

Reference

https://github.com/Tencent/ncnn

https://github.com/Z-Xiong/LightTrack-ncnn

https://github.com/FeiGeChuanShu/ncnn_Android_LightTrack