Multi-Stage VLM Pipeline for Zero-Shot Traffic Accident Understanding

• URL: http://arxiv.org/abs/2605.29325v1
• Date: Thu, 28 May 2026 04:00:26 GMT
• Title: Multi-Stage VLM Pipeline for Zero-Shot Traffic Accident Understanding
• Abstract summary: We present the 1st-place solution to the ACCIDENT challenge at the CVPR 2026 AUTOPILOT Workshop.<n>The challenge asks for zero-shot prediction of accident timing, impact centroid, and collision type from CCTV footage.<n>We build a three-stage pipeline (full-video joint prediction, time refinement, and single-frame grounding of the impact centroid), run the same pipeline a second time on a 235B Mixture-of-Experts sibling, blend the two outputs 9:1, and finally snap each predicted point onto the nearest vehicle detection.
• Score: 0.0
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We present the 1st-place solution to the ACCIDENT challenge at the CVPR 2026 AUTOPILOT Workshop, which asks for zero-shot prediction of accident timing, impact centroid, and collision type from CCTV footage. On a frozen Qwen3-VL-32B-Instruct checkpoint we build a three-stage pipeline (full-video joint prediction, time refinement, and single-frame grounding of the impact centroid), run the same pipeline a second time on a 235B Mixture-of-Experts sibling, blend the two outputs 9:1, and finally snap each predicted point onto the nearest vehicle detection. The final system reaches Public LB 0.55469 / Private LB 0.57080, roughly +0.21 over the strongest host baseline (Molmo-7B, 0.358) and wins the challenge. We ablate each component, report the negative results that shaped the final design, and release the code at https://github.com/fuumin621/cvpr2026-accident-1st-place-solution.

Related papers

• Zero-Shot Traffic Accident Detection via a Coarse-to-Fine VLM-Tracking Pipeline [0.10874100424278173]
  ACCIDENT @ CVPR benchmark evaluates exactly this joint prediction under a strict constraint.<n>We introduce a training-free, two-pass coarse-to-fine pipeline that pairs a frozen Qwen3-VL-32B-Instruct vision-language model with YOLO11x object detection and BoT-SORT tracking.<n>On the official 2,027-clip real-CCTV test set, our system achieves a three-way harmonic mean score of 0.504.
  arXiv  Detail & Related papers  (2026-08-09T19:08:56Z)
• Camera and LiDAR BEV Fusion for Cooperative 3D Object Detection on TUMTraf V2X [3.1508266388327324]
  We describe a Camera and LiDAR fusion detector developed for the TUMTraf V2X cooperative 3D object detection track of the DriveX 2026 challenge.<n>The detector fuses three roadside cameras with a fused infrastructure-plus-vehicle point cloud in a shared bird's-eye-view space.<n>The model reaches a 3D mAP of 0.85 on the publicbench Coda test split.
  arXiv  Detail & Related papers  (2026-06-11T07:14:10Z)
• A VideoMAE-v2 Approach to Zero-Shot Traffic Accident Anticipation [20.309082260025992]
  We propose a framework that bridges the gap between the frame-level temporal risk estimation task and coarsely labelled binary accident datasets.<n>Our method achieves 2nd place in the 2026 CVPR@AUTOPILOT Zero-Shot Accident Anticipation competition.
  arXiv  Detail & Related papers  (2026-06-08T14:25:16Z)
• Two-Pass Zero-Shot Temporal-Spatial Grounding of Rare Traffic Events in Surveillance Video [0.0]
  Grounding traffic accidents in real CCTV footage is a rare-event problem where training on labeled accident video is often prohibited.<n>We present a no-fine-tuning pipeline that elicits this joint output from frozen vision-language models.
  arXiv  Detail & Related papers  (2026-05-02T16:05:44Z)
• Instance-level Visual Active Tracking with Occlusion-Aware Planning [61.982298426203165]
  Visual Active Tracking (VAT) aims to control cameras to follow a target in 3D space.<n>VAT faces two key bottlenecks in real-world deployment: confusion from visually similar distractors and severe failure under occlusions.<n>We propose OA-VAT, a unified pipeline with three complementary modules.
  arXiv  Detail & Related papers  (2026-04-23T09:11:50Z)
• A Modular Zero-Shot Pipeline for Accident Detection, Localization, and Classification in Traffic Surveillance Video [0.0]
  We describe a zero-shot pipeline developed for the ACCIDENT @ CVPR 2026 challenge.<n>The challenge requires predicting when, where, and what type of traffic accident occurs in surveillance video, without labeled real-world training data.<n>Our method separates the problem into three independent modules.
  arXiv  Detail & Related papers  (2026-04-05T05:18:19Z)
• Collision-Aware Vision-Language Learning for End-to-End Driving with Multimodal Infraction Datasets [7.932761533792761]
  High infraction rates remain the primary bottleneck for end-to-end (E2E) autonomous driving.<n>We develop a Video-Language-Augmented Anomaly Detector (VLAAD)<n>VLAAD is a collision-aware plug-in module that can be seamlessly integrated into existing E2E driving models.
  arXiv  Detail & Related papers  (2026-03-26T22:32:52Z)
• LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixels [49.35636088613484]
  Joint Embedding Predictive Architectures (JEPAs) offer a compelling framework for learning world models in compact latent spaces.<n>We introduce LeModelWorld, the first JEPA that trains stably end-to-end from raw pixels using only two loss terms.<n>With 15M parameters trainable on a single GPU in a few hours, LeWM plans up to 48x faster than foundation-model-based world models.
  arXiv  Detail & Related papers  (2026-03-13T19:48:14Z)
• Collision Avoidance Detour for Multi-Agent Trajectory Forecasting [8.680676599607123]
  We present Collision Avoidance Detour (CAD), which won the 3rd place award in the 2023 Open dataset Challenge - Sim Agents. To satisfy the motion prediction factorization requirement, we partition all the valid objects into three mutually exclusive sets: Autonomous Driving Vehicle (ADV), World-tracks-to-predict, and World-others.
  arXiv  Detail & Related papers  (2023-06-20T16:05:24Z)
• CabiNet: Scaling Neural Collision Detection for Object Rearrangement with Procedural Scene Generation [54.68738348071891]
  We first generate over 650K cluttered scenes - orders of magnitude more than prior work - in diverse everyday environments. We render synthetic partial point clouds from this data and use it to train our CabiNet model architecture. CabiNet is a collision model that accepts object and scene point clouds, captured from a single-view depth observation.
  arXiv  Detail & Related papers  (2023-04-18T21:09:55Z)
• DeepAccident: A Motion and Accident Prediction Benchmark for V2X Autonomous Driving [76.29141888408265]
  We propose a large-scale dataset containing diverse accident scenarios that frequently occur in real-world driving. The proposed DeepAccident dataset includes 57K annotated frames and 285K annotated samples, approximately 7 times more than the large-scale nuScenes dataset.
  arXiv  Detail & Related papers  (2023-04-03T17:37:00Z)
• COPILOT: Human-Environment Collision Prediction and Localization from Egocentric Videos [62.34712951567793]
  The ability to forecast human-environment collisions from egocentric observations is vital to enable collision avoidance in applications such as VR, AR, and wearable assistive robotics. We introduce the challenging problem of predicting collisions in diverse environments from multi-view egocentric videos captured from body-mounted cameras. We propose a transformer-based model called COPILOT to perform collision prediction and localization simultaneously.
  arXiv  Detail & Related papers  (2022-10-04T17:49:23Z)
• Object Rearrangement Using Learned Implicit Collision Functions [61.90305371998561]
  We propose a learned collision model that accepts scene and query object point clouds and predicts collisions for 6DOF object poses within the scene. We leverage the learned collision model as part of a model predictive path integral (MPPI) policy in a tabletop rearrangement task. The learned model outperforms both traditional pipelines and learned ablations by 9.8% in accuracy on a dataset of simulated collision queries.
  arXiv  Detail & Related papers  (2020-11-21T05:36:06Z)
• Global Feature Aggregation for Accident Anticipation [8.57961305383434]
  We propose a novel Feature Aggregation (FA) block that refines each object's features by computing a weighted sum of the features of all objects in a frame. We use FA block along with Long Short Term Memory (LSTM) network to anticipate accidents in the video sequences.
  arXiv  Detail & Related papers  (2020-06-16T06:17:15Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.