Edubrief: NLP Powered NPTEL Video Summariser

This research paper presents a novel approach to summarizing NPTEL educational videos by applying NLP techniques and machine learning. The objective of this project is to develop a system that generates concise summaries of video content, extracts key frames, and derives text from diagrams within these frames. The system also generates questions based on the video content to enhance learning and comprehension. Our methodology incorporates several advanced tools and libraries. For the text summarization, we used NLTK, its stopwords, word tokenization, and sentence tokenization. We used the GUI, created with the Transformers library, which includes the pipeline, PlaintextParser, Tokenizer, and LexRankSummarizer for an efficient summarization. We used the spaCy library for question generation, renowned for its strong NLP features. The frame extraction is done using tkinter for the GUI, filedialog and messagebox for user interaction, and MoviePy for processing video. Finally, text extraction from diagrams makes use of TensorFlow’s ability to process images. The accompanying image illustrates the primary window of the system, which demonstrates the easy-to-use interface to interact with the summarization and question-generation features. This project has set ambitious goals for enhancing access and understanding of educational video content, providing a quality asset for both students and educators.