LLM 能否设计视频编码工具？以 Planar 模式为例的实证研究
Can LLMs Design Video Coding Tools? A Case Study on Planar Mode
AI 导读
一项 arXiv 实证研究探索 LLM 能否设计视频编码工具，以长期使用的帧内预测工具 Planar 模式为案例，采用“生成—评估”循环让 LLM 生成新的 Planar 预测器并依据编码评估反馈迭代改进。
正文 · 原文
Abstract:This paper explores whether large language models (LLMs) can design video coding tools, a highly challenging task due to the intricate algorithmic coupling of tool modifications. In particular, we present an empirical case study on the Planar mode, a long-standing intra prediction tool in video coding standards. Our experiments operate within a generation-and-evaluation loop, with the LLM generating new Planar predictors, encoder trials evaluating their coding performance, and the LLM re-generating refined implementations based on the evaluation feedback. We first examine directly replacing the default Planar mode in the Fraunhofer Versatile Video Encoder (VVenC) under its faster preset. Experimental results demonstrate that the LLM-generated mode can outperform the conventional Planar mode on this lightweight toolset, achieving 0.18% bitrate savings with 0.4% complexity overhead on the standard benchmark. We further extend our evaluation to the Enhanced Compression Model (ECM). Leveraging newly introduced directional Planar modes, we investigate two integration strategies: directly replacing them, and introducing the LLM-generated predictor as an additional prediction mode with new syntax elements. The empirical results suggest that both strategies can yield coding gains under a constrained low-resolution setting. Overall, this study offers preliminary evidence and practical insights, highlighting both the potential and open challenges of LLM-based coding tool design.
| Subjects: | Multimedia (cs.MM); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2609.01535 [cs.MM] |
| (or arXiv:2609.01535v2 [cs.MM] for this version) | |
| https://doi.org/10.48550/arXiv.2609.01535 arXiv-issued DOI via DataCite |
Submission history
From: Yingwen Zhang [view email]
[v1]
Tue, 1 Sep 2026 17:04:47 UTC (81 KB)
[v2]
Tue, 27 Oct 2026 02:50:07 UTC (81 KB)
来源：arXiv cs.AI · arxiv.org