TDBench: Benchmarking Vision Language Models on Top-Down Image Understanding

ICLR 2026 Conference Submission20420 Authors

Keywords: Benchmark, Evaluation, Trustworthiness, Large Language Models

TL;DR: This paper introduces TDBench, a new benchmark for top-down images that uses a novel rotational evaluation strategy to probe the reliability of Vision-Language Models.

Abstract: Top-down images play an important role in safety-critical settings such as autonomous navigation and aerial surveillance, where they provide holistic spatial information that front-view images cannot capture. Despite this, Vision Language Models (VLMs) are almost trained and evaluated on front-view benchmarks, leaving their performance in the top-down setting poorly understood. Existing evaluations also overlook a unique property of top-down images: their physical meaning is preserved under rotation. In addition, conventional accuracy metrics can be misleading, since they are often inflated by hallucinations or "lucky guesses", which obscures a model’s true reliability and its grounding in visual evidence. To address these issues, we introduce TDBench, a benchmark for top-down image understanding that includes 2000 curated questions for each rotation. We further propose RotationalEval (RE), which measures whether models provide consistent answers across four rotated views of the same scene, and we develop a reliability framework that separates genuine knowledge from chance. Finally, we conduct four case studies targeting underexplored real-world challenges. By combining rigorous evaluation with reliability metrics, TDBench not only benchmarks VLMs in top-down perception but also provides a new perspective on trustworthiness, guiding the development of more robust and grounded AI systems.

Primary Area: datasets and benchmarks

Submission Number: 20420