Research and Implementation of Robot Portrait Caricature Technology Based on Deep Learning

In the field of robot drawing portrait cartoon, the research and application of drawing technology is extremely critical, but there is the problem of inaccurate performance positioning. Common neural network algorithms fail to effectively solve the problem of inaccurate technical positioning when robots draw portraits, resulting in unsatisfactory results. Therefore, a research and implementation scheme of robot portrait cartoon technology based on deep learning algorithm is proposed, and the technology is evaluated. Firstly, the gradient descent theory is used to identify the influencing factors, and the indexes are divided according to the needs of drawing technology research and application, so as to reduce the interference factors in drawing technology research and application. Then, using gradient descent theory, a drawing technology research and implementation plan of deep learning algorithm is constructed, and the research results of drawing technology are comprehensively reviewed. MATLAB simulation results show that under specific evaluation conditions, the deep learning algorithm is superior to the standard neural network algorithm in the accuracy of mapping technology research and application and the time of influencing variables.