Answers are grounded in this paper and cite their sources.
The goal of prediction human motion is to analyze a subject's behaviors based on observed sequences and produced future body motions. In this work the deep neural network has been employed and proposed using wavelet transform with CNN-VAE model to analyze the input data to multi scales and extract features to encode it by CNN-VAE model, LSTM model has been used to predict encoded data and decoded it by used CNN-decoder to produce the new predicted frames. The propose system achieved best results in PSNR, MSE and SSIM and made the time of training and testing (prediction) faster. The experiments have been applied on two dataset: KTH and Weizmann and generate video of 1200 ms.
No takes yet. Share an insight, caveat, or question.
Ahmed et al. (2021) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: