With the continuous progress of science and technology, human beings have found that training machine learning can improve efficiency in many ways. Deep neural network is more widely used, such as face recognition, big data analysis and so on. In our daily life, people can sum up other people's emotions by observing other people's facial features, because different emotions will have different facial features, such as the shape of lips. Therefore, I choose facial expression recognition through deep neural network. In the experiment, I got the facial features by detecting the facial image, and trained the deep neural network through the existing training set, and test it.
Since the 21st century, artificial intelligence (AI) has developed rapidly, the rise of artificial intelligence has redefined the relationship between "human" and "computer". Scientists in the field of artificial intelligence are constantly trying to optimize computers so that they can better meet the needs of human beings, and even get closer to their emotional needs. Facial expression recognition is an important direction for computers to understand human emotions and an important aspect of human-computer interaction. Facial expression recognition refers to the selection of facial expressions from still photos or video sequences to determine the emotional and psychological changes of the characters. Based on TensorFlow, this report tries to solve the problem of facial expression recognition by using deep neural network (DNN). In this report, I use Dlib to preprocess the RaFD database, then build a DNN using TensorFlow. I use the processed database to train the model, and evaluate the accuracy of it.
Image Upscaling with Deep Convolutional Generative Networks
