Extremely Simple Multimodal Outlier Synthesis for Out-of-Distribution Detection and Segmentation

Researchers propose Feature Mixing, a novel method for multimodal out-of-distribution detection that achieves 10x to 370x speedup over existing approaches. The technique addresses safety-critical applications like autonomous driving by better detecting anomalous data across multiple sensor modalities.

• →Feature Mixing provides a simple and fast method for multimodal outlier synthesis with theoretical support for OOD detection.
• →The approach is modality-agnostic and applicable to various sensor combinations in real-world applications.
• →CARLA-OOD dataset introduces novel multimodal benchmarks for OOD segmentation across diverse conditions.
• →Method achieves state-of-the-art performance with 10x to 370x computational speedup over existing approaches.
• →Research addresses critical safety requirements for autonomous driving and robot-assisted surgery applications.