Multi-Center Backup Framework based on Audio and Video Complex Information Extraction Algorithm

This paper first introduces the principle, development and application of expert system, and proposes a robust agricultural science and education video text extraction method based on complex background. The algorithm mainly realizes video text extraction through video decoding, MSER text positioning, projection segmentation and Tesseract text recognition. An English-assisted teaching module based on expert system theory, which is aimed at teachers and users. The module classifies and summarizes English knowledge points using a combination of frame-based and production-based knowledge expressions. BP neural network is used to establish students’ autonomous learning and the system module analyzes the test situation of the students, enables the students to self-diagnose, and combines the memory forgetting curve through the students’ repeated practice.