GS-CLIP: Zero-shot 3D Anomaly Detection by Geometry-Aware Prompt and Synergistic View Representation Learning
This work addresses anomaly detection in 3D point clouds for applications like industrial inspection, where data scarcity and privacy are concerns, representing an incremental improvement over prior CLIP-based approaches.
The paper tackles zero-shot 3D anomaly detection by proposing GS-CLIP, a framework that uses geometry-aware prompts and synergistic view learning to address geometric detail loss and incomplete visual understanding in existing methods, achieving superior performance on four large-scale datasets.
Zero-shot 3D Anomaly Detection is an emerging task that aims to detect anomalies in a target dataset without any target training data, which is particularly important in scenarios constrained by sample scarcity and data privacy concerns. While current methods adapt CLIP by projecting 3D point clouds into 2D representations, they face challenges. The projection inherently loses some geometric details, and the reliance on a single 2D modality provides an incomplete visual understanding, limiting their ability to detect diverse anomaly types. To address these limitations, we propose the Geometry-Aware Prompt and Synergistic View Representation Learning (GS-CLIP) framework, which enables the model to identify geometric anomalies through a two-stage learning process. In stage 1, we dynamically generate text prompts embedded with 3D geometric priors. These prompts contain global shape context and local defect information distilled by our Geometric Defect Distillation Module (GDDM). In stage 2, we introduce Synergistic View Representation Learning architecture that processes rendered and depth images in parallel. A Synergistic Refinement Module (SRM) subsequently fuses the features of both streams, capitalizing on their complementary strengths. Comprehensive experimental results on four large-scale public datasets show that GS-CLIP achieves superior performance in detection. Code can be available at https://github.com/zhushengxinyue/GS-CLIP.