Content based lecture video retrieval

By the development of the internet technology and increasing internet providers have risen the amount of lecture videos as well as the other type of contents. While the impact of Covid 19 Pandemic around all over, that also changed the road map of education. Both the number of online educational content and the distance learning source and demand have increased alot. This rate of increase in the content and providers made it difficult to reach exact contents at its finest. The methods suggested in this study aim content based access to videos. Lecture videos have textual, audio and visual content. In order to illuminate this study, a lecture video dataset with 110 videos was created and the textual contents of the lecture videos in the Data Set were extracted by Optical Character Recognition (OCR) technology. Classification has done by three traditional machine learning methods and one deep learning method. Traditional machine learning methods applied in this study are Support Vector Machine, Naive Bayes and Random Forest methods. The deep learning method applied is the Long Short Term Memory method. In this study it is intended using machine learning and deep learning approaches to reach content based lecture videos.