AI-Based Video Editing using Stable Diffusion with Temporal Consistency Optimization

Read full text Summarise with VIRAT
Authors: Soham Pal
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This paper presents an AI-based video editing system that utilizes Stable Diffusion for transforming video frames based on textual prompts. The proposed approach converts an input video into individual frames, applies diffusion-based image-to-image translation, and reconstructs the edited frames into a final output video. To address the issue of temporal inconsistency, a smoothing technique is incorporated to reduce flickering and improve visual coherence across frames. The system is optimized for execution in resource-constrained environments by limiting frame count and reducing inference steps. Experimental results demonstrate the effectiveness of the proposed method in generating visually consistent and high-quality edited videos.

Fields of Study & Subjects

Generative Adversarial Networks and Image Synthesis Digital Media Forensic Detection Digital Media and Visual Art

Publication details

Journal / source
Figshare
Publication date
1 January 2026
Language
EN
Document type
Thesis