Emotion recognition by assisted learning with convolutional neural networks

Image emotion is the emotion hidden in or passed by a particular image. In this paper, a novel convolutional neural network is proposed to predict the emotion from an image. The proposed model consists of two parts: a binary positive-or-negative emotion classification network and a deep network for specific emotion recognition. During the network training, an assisted learning strategy is introduced to boost the recognition performance. Experimental results demonstrate that the proposed network is capable of extracting active level features and achieves significant gains in emotion recognition accuracy. (c) 2018 Elsevier B.V. All rights reserved.