VISTA: A Test-Time Self-Improving Video Generation Agent

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Long, Do Xuan, Wan, Xingchen, Nakhost, Hootan, Lee, Chen-Yu, Pfister, Tomas, Arık, Sercan Ö.
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917022295654400
author Long, Do Xuan
Wan, Xingchen
Nakhost, Hootan
Lee, Chen-Yu
Pfister, Tomas
Arık, Sercan Ö.
author_facet Long, Do Xuan
Wan, Xingchen
Nakhost, Hootan
Lee, Chen-Yu
Pfister, Tomas
Arık, Sercan Ö.
contents Despite rapid advances in text-to-video synthesis, generated video quality remains critically dependent on precise user prompts. Existing test-time optimization methods, successful in other domains, struggle with the multi-faceted nature of video. In this work, we introduce VISTA (Video Iterative Self-improvemenT Agent), a novel multi-agent system that autonomously improves video generation through refining prompts in an iterative loop. VISTA first decomposes a user idea into a structured temporal plan. After generation, the best video is identified through a robust pairwise tournament. This winning video is then critiqued by a trio of specialized agents focusing on visual, audio, and contextual fidelity. Finally, a reasoning agent synthesizes this feedback to introspectively rewrite and enhance the prompt for the next generation cycle. Experiments on single- and multi-scene video generation scenarios show that while prior methods yield inconsistent gains, VISTA consistently improves video quality and alignment with user intent, achieving up to 60% pairwise win rate against state-of-the-art baselines. Human evaluators concur, preferring VISTA outputs in 66.4% of comparisons.
format Preprint
id arxiv_https___arxiv_org_abs_2510_15831
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle VISTA: A Test-Time Self-Improving Video Generation Agent
Long, Do Xuan
Wan, Xingchen
Nakhost, Hootan
Lee, Chen-Yu
Pfister, Tomas
Arık, Sercan Ö.
Computer Vision and Pattern Recognition
Despite rapid advances in text-to-video synthesis, generated video quality remains critically dependent on precise user prompts. Existing test-time optimization methods, successful in other domains, struggle with the multi-faceted nature of video. In this work, we introduce VISTA (Video Iterative Self-improvemenT Agent), a novel multi-agent system that autonomously improves video generation through refining prompts in an iterative loop. VISTA first decomposes a user idea into a structured temporal plan. After generation, the best video is identified through a robust pairwise tournament. This winning video is then critiqued by a trio of specialized agents focusing on visual, audio, and contextual fidelity. Finally, a reasoning agent synthesizes this feedback to introspectively rewrite and enhance the prompt for the next generation cycle. Experiments on single- and multi-scene video generation scenarios show that while prior methods yield inconsistent gains, VISTA consistently improves video quality and alignment with user intent, achieving up to 60% pairwise win rate against state-of-the-art baselines. Human evaluators concur, preferring VISTA outputs in 66.4% of comparisons.
title VISTA: A Test-Time Self-Improving Video Generation Agent
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2510.15831