Online learning has revolutionised the way education is delivered and received. With the advent of technology, students can access information and education from anywhere in the world, at any time. However, one of the biggest concerns about online learning is that it is a one-way interaction, and there exists no real-time doubt clarification. In order to clear up doubts, students currently post their questions on forums and wait for an answer from the teacher. To surmount this challenge, a lecture video doubt clarification system is proposed. This system harnesses the potential of lecture videos and corresponding transcripts, incorporating domain knowledge and employing advanced techniques for the detection and recognition of lecture content. A unique aspect that sets the proposed system apart from the rest of the literature is its thorough investigation and comparison of both extractive and generative models. This paper proposes an approach for a doubt clarification system for e-learning videos, extensively discussing the above steps.