Tuesday, 3 November 2026 · 6 min read
Unified taxonomy and literature system organize agentic computer graphics research
This paper introduces a unified taxonomy for agentic computer graphics systems and an automated literature survey system that classifies over 300 papers, revealing trends in feedback use, update strategies, and resources across 2D and 3D graphics tasks.
Based on the paper Advances in Agentic Computer Graphics by Hengyu Liu, Jihyeon Je, Ruiqi Wang, Mingxuan Wu and 10 others
Challenges in organizing rapidly growing agentic graphics research
Computer graphics traditionally relies on code to create and manipulate visual content, but recent advances in large language models (LLMs) enable agents to generate graphics code and iteratively refine it using feedback from the environment. This agentic approach automates complex graphics tasks such as 3D asset creation, animation, and scene construction.
The rapid expansion of research in agentic graphics—over 300 papers by 2026—makes manual literature reviews difficult. Moreover, similar concepts are described with varying terminology across subfields like CAD, animation, and simulation, complicating systematic understanding and comparison.
To address these challenges, the paper aims to unify terminology and systematize the field by introducing a taxonomy that characterizes agentic graphics systems and by developing an automated literature survey system that classifies and analyzes papers at scale.
Sign in to read the whole paper
You have read the opening. The rest is free with a Bezen account, and no card is needed.
In the rest of this paper
- How the method works, step by step, with a flow chart
- Three figures from the paper, and how to read them
- How it was tested, what was found, and where it falls short