A deep learning approach for human face sentiment classification

This paper presents the development of a deep learning approach for human face sentiment classification. Bidirectional Long-Short Term Memory (Bi-LSTM) and recurrent neural networks (RNNs) are used for object-based segmentation in images, and CNN-RNN model is adopted for non-linear mapping. To test the applicability of this proposed approach, we have trained several deep neural networks to recognize facial expressions in 10,000 images. Our experiments show that the proposed approach can achieve an accuracy of 99.12% in classifying human face sentiment. Moreover, results indicate that the model not only boosts the classification model but also lessen the overhead.