Learning Discriminative Embedding for Hyperspectral Image Clustering Based on Set-to-Set and Sample-to-Sample Distances

• Published in: IEEE transactions on geoscience and remote sensing Vol. 58; no. 1; pp. 473 - 485
• Main Authors: Qin, Yao, Bruzzone, Lorenzo, Li, Biao
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.01.2020; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Affine hull (AH); Algorithms; Clustering; Clustering algorithms; Clustering methods; Computational modeling; Covariance matrix; Density; distance learning; Embedding; Entropy; Feature extraction; hyperspectral image (HSI); Hyperspectral imaging; Image classification; local covariance matrix representation (LCMR); Machine learning; Matrix representation; Multilayer perceptrons; remote sensing; sample-to-sample distance; set-to-set distance; siamese network; Training
• ISSN: 0196-2892; 1558-0644
• DOI: 10.1109/TGRS.2019.2937204

Recently, deep learning techniques have been introduced to address hyperspectral image (HSI) classification problems and have achieved the state-of-the-art performances. In this article, we propose a novel clustering algorithm for HSI based on learning embedding using the set-to-set and sample-to-sample distances (LSSDs). This technique consists of four main components: 1) oversegmentation; 2) generation of set-to-set and sample-to-sample distances; 3) learning embedding by training a siamese network; and 4) density-based spectral clustering. First, the HSI is oversegmented into superpixels by using the entropy rate superpixel (ERS) algorithm. Second, the set-to-set distances are obtained by representing the segmented sets of samples as affine hull (AH) models, whereas the sample-to-sample distances are computed by employing the local covariance matrix representation (LCMR) method. Third, sample pairs with the smallest and largest similarities are extracted according to the two distances. Then, these pairs are fed into the siamese multilayer perceptron (MLP) network and discriminative embeddings are learned by training the network with contrastive loss. Finally, density-based spectral clustering is applied to the deep embedding to obtain clustering results. Experimental results on three real HSIs demonstrate that the proposed method can achieve better performance than the considered baseline methods.