junxiant/person_reidentification

Person Re-identification

Findings

Using 4-Multi-Camera-Person-Tracking-and-Re-Identification.

Video result here: https://drive.google.com/file/d/1o8k0yR-L3YVaw_Hki3WNwiCzDu1DHN6C/view?usp=sharing

0-Multi-Camera-Object-Tracking-via-Transferring-Representation-to-Top-View

[https://github.com/Mhttx2016/Multi-Camera-Object-Tracking-via-Transferring-Representation-to-Top-View]

Multi view

1-Multi-Camera-Person-Re-Identification

[https://github.com/SurajDonthi/Multi-Camera-Person-Re-Identification]

No demo script?

Multi view

Under issues the author said "Currently, only the metrics are stored. However, feel free to create a pull request to add storing of the test outputs."

2-person-detect-reidentification

[https://github.com/yas-sim/person-detect-reidentification]

Can be used straight out of the box.

One script to run the code (But need to modify to accept video input).

Also has person/face re-id.

But requires OpenVINO installation.

3-Person-Re-identification

[https://github.com/Nirvan101/Person-Re-identification]

Easy to use

But re-id model is from 2015 [https://www.cv-foundation.org/openaccess/content_cvpr_2015/papers/Ahmed_An_Improved_Deep_2015_CVPR_paper.pdf]

Uses tensorflow 1....

4-Multi-Camera-Person-Tracking-and-Re-Identification

[https://github.com/samihormi/Multi-Camera-Person-Tracking-and-Re-Identification] Easy to use

But uses yolov3/v4

Uses tf 1 also...

192 frames took 511 seconds (CPU)

1 frame = 2.66 seconds

Video result here: https://drive.google.com/file/d/1o8k0yR-L3YVaw_Hki3WNwiCzDu1DHN6C/view?usp=sharing

Follow the README.md in this folder and run python demo.py --videos ../T3.mp4 --version v4

Video clip taken from:

https://www.youtube.com/watch?v=__eLCXUKtec