A Research of Character Recognition Based on Residual Neural Network

摘要

Aiming at the problem of video person recognition, this paper analyzes current technical environment and current technical problems, and designs a set of character recognition model based on residual neural network. It extracts key frames, locates characters, extracts characters of video and makes similar features with target characters. Comparison, to determine the identity of the person, and to explore the practicality of the model, through experiments, the identity of the person has achieved very good experimental results.