IR3D-Bench: Evaluating Vision-Language Model Scene Understanding as Agentic Inverse Rendering
This addresses the need for better evaluation of scene understanding in vision-language models, though it is incremental as it builds on existing analysis-by-synthesis paradigms.
The paper tackles the problem of evaluating whether vision-language models truly understand scenes by introducing IR3D-Bench, a benchmark that challenges them to actively recreate 3D structures from images using tools, with initial experiments showing limitations in visual precision.
Vision-language models (VLMs) excel at descriptive tasks, but whether they truly understand scenes from visual observations remains uncertain. We introduce IR3D-Bench, a benchmark challenging VLMs to demonstrate understanding through active creation rather than passive recognition. Grounded in the analysis-by-synthesis paradigm, IR3D-Bench tasks Vision-Language Agents (VLAs) with actively using programming and rendering tools to recreate the underlying 3D structure of an input image, achieving agentic inverse rendering through tool use. This "understanding-by-creating" approach probes the tool-using generative capacity of VLAs, moving beyond the descriptive or conversational capacity measured by traditional scene understanding benchmarks. We provide a comprehensive suite of metrics to evaluate geometric accuracy, spatial relations, appearance attributes, and overall plausibility. Initial experiments on agentic inverse rendering powered by various state-of-the-art VLMs highlight current limitations, particularly in visual precision rather than basic tool usage. IR3D-Bench, including data and evaluation protocols, is released to facilitate systematic study and development of tool-using VLAs towards genuine scene understanding by creating.