VeriScene reconstructs crime scenes accurately from mixed legal evidence
VeriScene: Reconstructing Crime Scenes from Legal Evidence via World-Model Agent
Computer Vision and Pattern RecognitionCryptography and Security
Summary
It can be hard to piece together exactly what happened at a crime scene using a lot of different evidence like photos and witness accounts, especially because some stories may contradict or be unreliable. The authors created VeriScene, a system that carefully combines these pieces of evidence to build a realistic, physics-based re-creation of the crime scene. It checks its own work to make sure the story it tells matches the evidence and that the movements and events follow the laws of physics. In tests with fake crime cases, VeriScene was better at including all evidence and staying true to the facts than other similar methods.
world modelmultimodal inputforensic evidencecrime scene reconstructionlegal evidence fusionphysical plausibilitywitness statementsfactual consistencykeyframe renderingLLM orchestration
Authors
Kevin Chuanpu Fu, Yongsen Zheng, Zee Kin Yeong, Kwok-Yan Lam
Abstract
World models take multimodal inputs like text, photos, and diagrams to generate dynamic scenes in accordance with the laws of physics, thus opening a compelling application: fusing multimodal legal evidence to re-create a crime scene and re-enact how an offence could have been committed. However, feeding the raw, unorganized evidence into a world model fails in forensic use: it silently drops evidence, glosses over contradictory testimony, and produces motion that violates the evidentiary record. This paper presents VeriScene, an agent that orchestrates the world model: it reconstructs crime scenes from forensic photographs and witness statements of varying reliability, keeping every claim traceable to evidence and every motion physically plausible. VeriScene iteratively fuses the evidence into a cited narrative under an auditing loop, verifies the hypothesized dynamics via probe rollouts in the world model with corrective constraint injection, and renders the offence as a re-enactment video from a fused keyframe. On a benchmark of 25 crime scenarios across 7 physically-driven case types (139 forensic-style photographs and 65 statements with planted unreliability), VeriScene attains 0.9014 evidence coverage and 0.7217 factual consistency (0-1 scale) on the 20 test scenes, outperforming an end-to-end multimodal-LLM baseline by 20.35% in factual consistency and 34.88% in temporal coherence, while generalizing across four LLM orchestration backends at USD 1.82 per scene.