MulSen-AD Dataset
Multi-Sensor Object Anomaly Detection: Unifying Appearance, Geometry, and Internal Properties
READ MORE
PAPER
Multi-Sensor Object Anomaly Detection: Unifying Appearance, Geometry, and Internal Properties
Wenqiao Li1*, Bozhong Zheng1*, Xiaohao Xu2*, Jinye Gan1, Fading Lu1, Xiang Li1, Na Ni1, Zheng Tian1, Shenghua Gao3, Yingna Wu1
1ShanghaiTech University, 2University of Michigan, Ann Arbor, 3The University of Hong Kong
Abstract

Object anomaly detection is essential for industrial quality inspection, yet traditional single-sensor methods face critical limitations. They fail to capture the wide range of anomaly types, as single sensors are often constrained to either external appearance, geometric structure, or internal properties. To overcome these challenges, we introduce MulSen-AD, the first high-resolution, multi-sensor anomaly detection dataset tailored for industrial applications. MulSen-AD unifies data from RGB cameras, laser scanners, and lock-in infrared thermography, effectively capturing external appearance, geometric deformations, and internal defects. The dataset spans 15 industrial products with diverse, real-world anomalies. We also present MulSen-AD Bench, a benchmark designed to evaluate multi-sensor methods, and propose MulSen-TripleAD, a decision-level fusion algorithm that integrates these three modalities for robust, unsupervised object anomaly detection. Our experiments demonstrate that multi-sensor fusion substantially outperforms single-sensor approaches, achieving 96.1% AUROC in object-level detection accuracy. These results highlight the importance of integrating multi-sensor data for comprehensive industrial anomaly detection. The dataset and code are publicly available to support further research.

The Dataset
NORMAL AND ABNORMAL SAMPLES/div>
Collection Pipeline
Fifteen Object Classes
Anomaly Detection Results on the Mulsen-AD Dataset (Object-Level AUROC)
Category Single Double Triple
RGB Infrared PC RGB+PC Infrared+PC RGB+Infrared RGB+Infrared+PC
Capsule0.9520.8960.8170.9520.8980.9770.967
Cotton0.8680.9220.8350.8830.9380.9580.960
Cube0.9490.9680.4230.9490.9680.9680.980
Spring Pad0.8710.7790.8170.9040.7920.8170.879
Screw1.0000.9350.6561.0000.9480.9741.000
Screen0.7590.9910.7560.7810.9910.9810.938
Piggy1.0000.9600.4001.0000.9601.0001.000
Nut0.9760.5900.5410.9550.5900.7690.959
Flat pad0.7100.9500.8300.7800.9500.9370.863
Plastic Cylinder0.8711.0000.6000.8711.0001.0000.993
Zipper0.9580.9910.8070.9700.9940.9880.994
Button Cell1.0000.8260.8131.0000.8320.9771.000
Toothbrush0.9510.9200.9200.9660.9240.9390.955
Solar Panel0.8260.9410.4000.8030.9330.9770.949
Light0.9670.9690.4030.9690.9690.9610.972
Mean0.9110.9090.6680.9190.9120.9480.961