Artificial Intelligence Education Platform for Classrooms Based on Action Interaction

To address the challenge of limited feedback in the learning process, hindering users from discerning deviations from standard actions, the author suggests a design framework for a distributed action teaching platform leveraging pose estimation algorithms. This innovative platform integrates artificial intelligence and distributed technology. The development is structured with a clear separation between the front-end and back-end, harnessing distributed and cluster technologies to effectively tackle high concurrency and ensure optimal availability. The platform uses the Tensorflow.js framework to perform real-time action recognition, analysis, and guidance in the front-end, avoiding issues such as privacy leakage and server overload that are prone to occur during data exchange between the front-end and back-end. The experimental results showed that by analyzing the user’s past course learning, Echarts was used to generate statistical charts and graphical exercise reports. The highest number of exercises on Saturday was 7, and the graphical exercise status showed that the right hand learning ability, left leg learning ability, left hand learning ability, and right leg learning ability were 84, 72, 59, and 39, respectively. Conclusion: By scoring and comparing algorithms, users can continuously correct their actions during use, achieving positive feedback effects in action learning.