AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

CrossView: Can Vision-Language Models Reason Across Cameras?

Sahil Shah, S P Sharan, Harsh Goel, Manvik Pasula, Adithya Hebbalae, Minkyu Choi, Sandeep Chinchali

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: VLM capability analysis · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
scale up model or use proprietary GPT-5.2 on multi-camera questionsVLMs on CrossView · CrossViewaccuracy: GPT-5.2 25.4 on nuScenes temporal; 34.6 on Ego-Exo4D best cameraNo clear effect
use all six cameras instead of the front camera only (ablation)Qwen2.5-7B · CrossView nuScenes countingaccuracy (%): 18.7 -> 27.5Improved
use all six cameras instead of the front camera only (ablation)Qwen2.5-7B · CrossView nuScenes spatialaccuracy (%): 41.5 -> 45.9Improved
use all six cameras instead of the front camera only (ablation)Qwen2.5-7B · CrossView nuScenes event orderingaccuracy (%): 32.0 (front) vs 24.4 (all cameras)Hurt
use all cameras instead of ego or best exo camera (ablation)Qwen2.5-7B · CrossView Ego-Exo4D temporalaccuracy (%): 40.4-40.7 -> 45.2Improved
stitched tiling instead of uniform per-camera sampling (ablation)Qwen2.5-7B · CrossView nuScenesaccuracy (%): counting 27.5 -> 38.0; event ordering 24.4 -> 39.6Improved
stitched tiling instead of uniform per-camera sampling (ablation)Qwen2.5-7B · CrossView MEVAaccuracy (%): spatial 17.6 -> 34.2; event ordering 83.2 -> 70.9Mixed
stitched tiling instead of uniform per-camera sampling (ablation)Qwen2.5-7B · CrossView AgiBotaccuracy (%): event ordering 54.4 -> 48.4Hurt
stitched tiling instead of uniform sampling (ablation)Qwen2.5-7B · CrossView Ego-Exo4Daccuracy (%): temporal 45.2 -> 53.6; event ordering 49.2 -> 48.0Mixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy