Work

CodeVisionary: An Automated Visual Explanation System for Coding Problems

Year: 2026
Type: conference-paper
Abstract: Algorithm based problem solving is a difficult process for self-taught learners, especially when the explanations only rely on lectures and textual mode. There will be mistakes in written solutions and also hard for understanding. In this paper, we present CodeVisionary, an automated system that creates narrated tutorial videos from algorithmic problem descriptions. The system combines large language models for reasoning, neural text-to-speech for narration, and automated slide generation to explain how algorithms work visually. CodeVisionary serves as a complete end-to-end process that includes understanding the problem, breaking down the solution, creating audio, generating visual slides, and rendering the video. Our experiments showed that the system produces tutorial videos efficiently while keeping good synchronization between the narration and visuals. Feedback from undergraduate learners indicates that these videos improve their understanding of concepts and help them grasp key algorithmic ideas more quickly than traditional text-based explanations.
Language: English
Open Access status: hybrid
OpenAlex ID: W7213423589
FWCI:
Cites: 12
Cited by:
Related to: 10