OoDDINO:A Multi-level Framework for Anomaly Segmentation on Complex Road Scenes

2025-07-02

AI Summary

Existing pixel-level anomaly segmentation methods suffer from two key limitations: (1) neglecting pixel-wise spatial correlations, leading to fragmented segmentations, and (2) employing global thresholds ill-suited to foreground/background heterogeneity, causing false positives and missed detections. To address these challenges in complex road scenes, we propose a coarse-to-fine multi-stage detection framework. First, object-level localization mitigates fragmentation; second, an orthogonal uncertainty-aware fusion strategy enhances robustness in anomaly localization. Furthermore, we introduce an adaptive dual-threshold network that enables discriminative pixel-wise segmentation for foreground and background regions. Our framework adopts an uncertainty-guided cascaded architecture, supporting plug-and-play integration of diverse feature representations and uncertainty estimators. Evaluated on two benchmark datasets, our method achieves significant improvements over state-of-the-art approaches, demonstrating superior accuracy, fine-grained segmentation capability, and model compatibility.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Uncertainty Representations

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Anomaly segmentation aims to identify Out-of-Distribution (OoD) anomalous objects within images. Existing pixel-wise methods typically assign anomaly scores individually and employ a global thresholding strategy to segment anomalies. Despite their effectiveness, these approaches encounter significant challenges in real-world applications: (1) neglecting spatial correlations among pixels within the same object, resulting in fragmented segmentation; (2) variabil ity in anomaly score distributions across image regions, causing global thresholds to either generate false positives in background areas or miss segments of anomalous objects. In this work, we introduce OoDDINO, a novel multi-level anomaly segmentation framework designed to address these limitations through a coarse-to-fine anomaly detection strategy. OoDDINO combines an uncertainty-guided anomaly detection model with a pixel-level segmentation model within a two-stage cascade architecture. Initially, we propose an Orthogonal Uncertainty-Aware Fusion Strategy (OUAFS) that sequentially integrates multiple uncertainty metrics with visual representations, employing orthogonal constraints to strengthen the detection model's capacity for localizing anomalous regions accurately. Subsequently, we develop an Adaptive Dual-Threshold Network (ADT-Net), which dynamically generates region-specific thresholds based on object-level detection outputs and pixel-wise anomaly scores. This approach allows for distinct thresholding strategies within foreground and background areas, achieving fine-grained anomaly segmentation. The proposed framework is compatible with other pixel-wise anomaly detection models, which acts as a plug-in to boost the performance. Extensive experiments on two benchmark datasets validate our framework's superiority and compatibility over state-of-the-art methods.

Problem

• Fragmented segmentation due to ignored pixel spatial correlations
• Global thresholds causing false positives or missed anomalies
• Need for dynamic region-specific anomaly segmentation thresholds

Innovation

• Multi-level framework for anomaly segmentation
• Uncertainty-guided anomaly detection model
• Adaptive dual-threshold network for segmentation