Structured SVM backpropagation to Convolutional Neural Network applying to Human Pose Estimation

Peerajak Witoonchart, Prabhas Chongstitvatana

Department of Computer Engineering, Chulalongkorn University

Neural Networks (2017)
http://dx.doi.org/10.1016/j.neunet.2017.02.005

Abstract

In this work, we show, for the first time, how to formulate Structured SVM as two layers of Convolutional Neural Network, the top layer of which is loss augmented inference layer, and the bottom is normal convolutional layer. We show that Deformable Part Model can be learned with newly created Structured SVM neural network by propagating the error of Deformable Part Model back propagate to Convolutional Neural Network. The forward propagation calculate loss augmented inference. The back propagation calculate the gradient from the loss augmented inference layer to convolutional layer. By doing so, we create a new type of convolutional neural network: Structured SVM Convolutional Neural Network, which is then applied to Human Pose Estimation problem. This new creation is a neural network, which can be used as the last layers of deep learning. Our method jointly learn structural model parameters and appearance model parameters. We implement our method as a new layer of existing Caffe library. Our source code is available for download. 

Keywords: Structured SVM, Convolutional Neural Network, Back Propagation, Deformable Part Model, Human Pose Estimation