ABSTRACT
The significance of video analytics in online conferencing has grown quite a lot due to the drastic increase in the usage of online conferencing platforms in recent years. Deep learning techniques have shown potential across various applications within video analytics, including tasks such as object detection, scene classification, and event recognition. This review paper takes a look at a comprehensive overview of the research on the use of deep learning for video analytics in the context of online conferencing. The paper summarizes the various approaches used for video analytics, including convolutional neural networks, various machine learning models, and other techniques, and evaluates their performance on a range of datasets and usecases. The review also shows the challenges and limitations associated with these methods, including scalability and variability in accuracy, and the need for further research in these areas. The paper concludes by discussing the future potential of deep learning for video analytics on online conferencing and the potential for new and innovative approaches to emerge in the future.
