Artificial Intelligence based Approach for Virtual Character Recognition

Virtual teaching involves providing educational instruction and content to students using digital methods, predominantly over the internet. In virtual teaching, educators and learners engage in remote interactions, often from separate physical locations, utilizing a range of online tools and technologies. This approach can include real-time, synchronous sessions through video conferencing, as well as asynchronous learning, where students access pre-recorded lectures, assignments, and resources at their own convenience. Here, artificial intelligence techniques such as Convolutional Neural Network (CNN) and Improved Recurrent Neural Network (IRNN) are used for character recognition in virtual teaching. Initially, some virtual teaching videos are collected and it split into frames. Each frame is pre-processed by removing noise using auto-encoder. The de-noised images are given as input to CNN and IRNN for character recognition in virtual teaching videos. The experimental findings confirm that, for character recognition in virtual teaching, the IRNN demonstrates 5.88%, 4.6%, 4.65% and 5.78% better accuracy, precision, recall and F-measure compared to the CNN.