Joint learning framework of superpixel generation and fuzzy sparse subspace clustering for color image segmentation
Chengmao Wu,Jingtian Zhao
DOI: https://doi.org/10.1016/j.sigpro.2024.109515
IF: 4.729
2024-04-25
Signal Processing
Abstract:Sparse subspace clustering (SSC) is an important image segmentation method that constructs a self-representation coefficient matrix to represent the relationships between pixels, and then uses spectral clustering to achieve clustering. Compared with other unsupervised segmentation algorithms, SSC has good segmentation performance. However, SSC has a high computational complexity when processing large-sized image. To improve computational efficiency, many researchers have proposed superpixel-based SSC algorithms, which process superpixels instead of pixels and improve efficiency through preprocessing. Due to the sensitivity of superpixel generation to noise, superpixel-based SSC algorithms still have poor robustness. Additionally, preprocessing increases the complexity of the algorithm. To address these issues, this paper proposes a robust superpixel-based fuzzy sparse subspace clustering algorithm. This algorithm combines fuzzy sparse subspace clustering with superpixel generation, and it constructs a unified optimization learning framework through fuzzy C-multiple-means clustering to improve segmentation performance and reduce complexity. Additionally, this paper introduces additional features of superpixel in sparse subspace clustering to further enhance the segmentation performance of the algorithm. Experiment results indicate that the proposed algorithm not only outperforms existing state-of-the-art robust segmentation algorithms independent of superpixels, but also is superior to the latest superpixel-based segmentation algorithms.
engineering, electrical & electronic