Create GAN to detect anomalies from human poses

***** Pls only bid if you know this stuff and are confident that you can complete it ****

I want to create a network for human pose anomaly detection using GANs (or any other method )

Problem statement:

Given n frames (32 or 64), track and extract human pose and detect whether it's an anomaly or not.

Reference: https://github.com/arunppsg/TadGAN

I can provide dataset for the task,