Human Activity Understanding

Understanding human activity is very challenging even with recently developed 3D/depth sensors. To solve this problem, this work investigates a novel deep structured model that adaptively decomposes an activity into temporal parts using convolutional neural networks (CNNs). The proposed model advances two aspects of the traditional deep learning approaches. First, a latent temporal structure is introduced into the deep model, accounting for large temporal variations in diverse human activities. In particular, we utilize the latent variables to decompose the input activity into a number of temporally segmented subactivities and feed them into the parts (ie, subnetworks) of the deep architecture. Second, we incorporate a radius-margin bound as a regularization term into our deep model, which effectively improves the generalization performance for classification (Reprinted by permission from Springer Nature …