Real-IAD D³ is a high-precision multimodal dataset jointly released by Shanghai Jiao Tong University, Shanghai Ocean University and other institutions. The relevant paper results are:Real-IAD D3: A Real-World 2D/Pseudo-3D/3D Dataset for Industrial Anomaly Detection", and has been included in the top computer vision conference CVPR 2025.

The dataset contains 20 industrial product categories, 69 defect types, and a total of 8,450 samples, including 5,000 normal samples and 3,450 abnormal samples. The data is derived from a real production line, including material preparation, defect manufacturing, image acquisition settings, marking and cleaning, annotation and other steps. Each sample contains synchronized RGB images, pseudo 3D photometric stereo images, and 3D point cloud data with micron-level accuracy, providing richer information for anomaly detection.

Data Collection Process