Gesture recognition based on CNN and DCGAN for calculation and text output

Abstract

In the past few years, with the continuous improvement of hardware conditions, deep learning had performed well in solving many problems, such as visual recognition, speech recognition, and natural language processing. In recent years, human-computer interaction behavior has appeared more and more in daily life. Especially with the rapid development of computer vision technology, the human-centered human-computer interaction technology is bound to replace computer-centered human-computer interaction technology. The study of gesture recognition is in line with this trend, and gesture recognition provides a way for many devices to interact with humans. The traditional gesture recognition method requires manual extraction of feature values, which is a time-consuming and laborious method. In order to break through the bottleneck, we propose a new gesture recognition algorithm based on the convolutional neural network and deep convolution generative adversarial networks. We apply this method to expression recognition, calculation, and text output, and achieve good results. The experiments show that the proposed method can train the model to identify with fewer samples and achieve better gesture classification and detection effects. Moreover, this gesture recognition method is less susceptible to illumination and background interference. It also can achieve an efficient real-time recognition effect.