GAN-motion-Prediction

An LSTM based GAN for motion synthesis has a few issues reading H3.6M data from A.Jain et al , will fix soon.

Prediction of the human motion model has been an intrinsic part of several applications over rampant fields like gaming, augmented reality and cinematic graphics.The ability to estimate motion, ahead of time, helps robots predict the human action and thus reduce time to react effectively. In real time applications such as pedestrian motion prediction,the availability of long motion sequences is tenuous. In this work, we propose a new architecture to model the human motion from noise.We utilize the data synthesizing ability of the Generative Adversarial Networks(GANs) to provide motion prediction that works with an LSTM-RNN foundation. The successful Recurrent Neural Network is used as a discriminator in training a weaker LSTM generator that we later exploit in creating ground truth like data from random noise.Pivoting on the evaluation wonts used in recent works,We collate the recent motion prediction techniques and compare the results.We also evaluate the training procedures,input requirements and complexity of the structures,thus illustrating the simplicity and accuracy of a GAN based prediction model.