Research and publications

AcciSAGE

Spatio-Temporal Accident Grounding in Surveillance Videos via Error-Aware VLM Tuning

Master’s research · Manuscript · 2026 · Qwen3-VL, LoRA, SAPO, Computer Vision

Error-aware vision-language model tuning for temporal, spatial, and collision-type grounding in surveillance accident videos.

Research introduction video

A research introduction showing global-to-local accident grounding, task-specific LoRA adapters, SAPO refinement, and accident case demonstrations.

AcciSAGE · 1:51 · English on-screen text and subtitles

Open the MP4 video · Download video

Abstract

AcciSAGE is a global-to-local framework for spatio-temporal accident grounding in surveillance videos. It uses one open-weight vision-language backbone with task-specific LoRA adapters for temporal, spatial, and collision-type predictions. Supervised fine-tuning (SFT) is followed by error-aware SAPO refinement with task-aligned rewards.

AcciSAGE method overview diagram
Global-to-local inference and SFT-to-SAPO refinement. Source: AcciSAGE. Select the image to view at full size.

Method

Results

Reference results from the public repository on the ACCIDENT IID test set. T₁ is the Gaussian temporal grounding score at σ = 1 second; S is the spatial score, C is collision-type accuracy, and ACCS is their harmonic mean.

Reference results on the ACCIDENT IID test set
MetricSFT-onlyAcciSAGE (SAPO)
T₁0.55630.5672
S0.52360.5352
C0.53820.5493
ACCS0.53900.5503

Source: AcciSAGE reference results; evaluation protocol.

Research status and thesis record

Master’s research, 2026. Manuscript with a public reference implementation.

Master’s thesis: Spatio-Temporal Traffic Accident Grounding in Surveillance Videos via Error-Aware VLM Tuning. National Taipei University of Technology, 2026. Record in Taiwan’s National Digital Library of Theses and Dissertations.

Links

[Master’s thesis] [Source code]

Contact: cjz89k@gmail.com