FACIAL EXPRESSION RECOGNITION METHOD BASED ON CONVOLUTIONAL NEURAL NETWORK

摘要

Provided is a facial expression recognition method based on a convolutional neural network. The facial expression recognition method according to an embodiment of the present invention includes the steps of: detecting a facial region; generating a reference image from the detected facial region; generating a plurality of facial images from the reference image; and recognizing an expression by learning the images in a neural network. Accordingly, in the facial expression recognition, it is possible to minimize an over-fitting problem that occurs when a small amount of images are learned in a deep-layer neural network by performing learning based on the convolutional neural network after data expansion.;COPYRIGHT KIPO 2018