A Unified Model Selection Technique for Spectral Clustering Based Motion Segmentation
This work addresses a practical limitation in motion segmentation for applications like robotics and autonomous driving, but it is incremental as it combines existing model selection techniques.
The paper tackles the problem of motion segmentation requiring prior knowledge of the number of motions by proposing a unified model selection technique to automatically infer this number, achieving competitive results on the KT3DMoSeg dataset compared to baselines with ground truth cluster counts.
Motion segmentation is a fundamental problem in computer vision and is crucial in various applications such as robotics, autonomous driving and action recognition. Recently, spectral clustering based methods have shown impressive results on motion segmentation in dynamic environments. These methods perform spectral clustering on motion affinity matrices to cluster objects or point trajectories in the scene into different motion groups. However, existing methods often need the number of motions present in the scene to be known, which significantly reduces their practicality. In this paper, we propose a unified model selection technique to automatically infer the number of motion groups for spectral clustering based motion segmentation methods by combining different existing model selection techniques together. We evaluate our method on the KT3DMoSeg dataset and achieve competitve results comparing to the baseline where the number of clusters is given as ground truth information.