Random Walk on Pixel Manifolds for Anomaly Segmentation of Complex Driving Scenes
This work addresses anomaly segmentation for autonomous driving systems, offering an incremental improvement by refining embeddings to handle scene diversity.
The paper tackles the problem of anomaly segmentation in complex driving scenes by addressing distorted pixel embedding manifolds that hinder accurate anomaly scoring, proposing Random Walk on Pixel Manifolds (RWPM) to refine embeddings and improve performance, with experiments showing it achieves state-of-the-art results.
In anomaly segmentation for complex driving scenes, state-of-the-art approaches utilize anomaly scoring functions to calculate anomaly scores. For these functions, accurately predicting the logits of inlier classes for each pixel is crucial for precisely inferring the anomaly score. However, in real-world driving scenarios, the diversity of scenes often results in distorted manifolds of pixel embeddings in the space. This effect is not conducive to directly using the pixel embeddings for the logit prediction during inference, a concern overlooked by existing methods. To address this problem, we propose a novel method called Random Walk on Pixel Manifolds (RWPM). RWPM utilizes random walks to reveal the intrinsic relationships among pixels to refine the pixel embeddings. The refined pixel embeddings alleviate the distortion of manifolds, improving the accuracy of anomaly scores. Our extensive experiments show that RWPM consistently improve the performance of the existing anomaly segmentation methods and achieve the best results. Code is available at: \url{https://github.com/ZelongZeng/RWPM}.