Introduction - If you have any usage issues, please Google them yourself
This paper presents a number of semi-supervised learning-based automatic video annotation methods. Through several common semi-supervised learning methods, such as self-training, cross training, and Co an analysis of EM and other methods for them (mainly from the training and cross-training methods) in the video annotation applications, the limitations in enhancing the classification accuracy and model updating have done a thorough study and submit a corresponding improvement measures.