Classroom learning behavior recognition method for English teaching students based on adaptive feature fusion

A new method of English teaching students' classroom learning behaviour recognition based on adaptive feature fusion is proposed aiming at the problem of low recognition rate of classroom learning behaviour recognition. First, the video images of English teaching class were collected and then divided into frames and grey-scale processing. Secondly, the improved guided filtering algorithm was used to enhance the image. Then, the maximum inter-class variance method was used to segment the image. Finally, SIFT algorithm was introduced to design an adaptive feature fusion architecture, which adaptively allocates feature weights and fuses shallow and deep features to realise learning behaviour recognition. The experimental results show that the proposed method has a peak signal-to-noise ratio of 51.7 dB, a recognition rate of 97.9%, and a maximum delay of 1.9 s, which can accurately identify classroom learning behaviour.