A Comparative Study of Video Watermarking using Various Machine Learning Algorithms

Watermarking refers to the process of inserting data or information into your content for the purpose of identification or protection to maintain its authenticity and copyrights. Watermarking is like stenography and is used in many places such as in photos, videos, money, and stamp to assist in identifying counterfeiting. There are many techniques developed for watermarking in which some are wavelet based and some are transform based. Watermarking is a two-way process in which first the watermark is embedded over the data by using embedding techniques such as two-way dimensional discrete wave transform and watermark extraction techniques such as the extreme learning machines, neural networks etc. To provide the clear overview of the existing algorithms in combination with extreme learning machine we went through existing studies from 2014-2022. This review gives an exhaustive study of existing methods and suggests some recommendations for further enhancements to be done by using the various deep learning and machine learning techniques.