Today, in the academic, corporate and health care world video streaming is widely used to deliver presentations, lectures and to perform remote diagnosis. These videos contain a variety of information presented in various media. For example, a lecture video consists of information presented on assorted media such as a computer and a white board. Thus, to refer back to one key point one has to manually browse through the parts of the video which is inefficient and time consuming. Documenting the information provided in the video will not only reduce this problem but also summarizes the video without going through all the video details. This project is based on previous work on annotating the video by extracting text from it through the use of prior information collected from the user in order to accurately segment the text.