Generalized Robust PCA: A New Distance Metric Method for Underwater Target Recognition

Abstract

Inspired by the importance of distance metrics and the structure-preserving ability of features, a novel recognition method for underwater targets, called generalized robust principal component analysis (GRPCA), is proposed in this paper. Several advantages of GRPCA are summarized as follows. First, GRPCA employs the l 2,p -norm as the distance metric for calculating the reconstruction error and variance of projected data and attempts to minimize the sum of the ratios between the reconstruction error and the variance for each data sample. This approach allows it to extract the feature information of an image more accurately, which is important for recognition and representation. Second, the proposed GRPCA algorithm not only is robust but also retains the desirable properties of PCA, such as rotational invariance. Moreover, we present a simple yet efficient iterative update algorithm to solve the GRPCA problem. Finally, on the basis of GRPCA, underwater target recognition technology is developed. The extensive experiments on several underwater optical image databases show that our method is more effective and advantageous than other subspace learning algorithms are.INDEX TERMS Distance metric, generalized robust principal component analysis (GRPCA), underwater target recognition.

Application of Locally Invariant Robust PCA for Underwater Image Recognition

“…In this section, we validate the effectiveness of the proposed method using the FDT-UT [43], JEDI-UT [44] and EPIDHEU-UT [38] databases and compare it with state-ofthe-art image recognition algorithms (PCA [8], PCA-L1 greedy [22], R1-PCA [34], Angle PCA [37], 2, p l -PCA ( = 0.5 p ) [15] and PgLPCA [40]. FDT-UT and JEDI-UT contain relatively well-known underwater image information.…”

“…To address these problems, Ding et al [34] presented a novel unsupervised feature selection algorithm called R1-PCA, which employs an 1 l -norm with rotation invariance (R1-norm) to minimize the reconstruction error. Motivated by the R1-PCA method, a series of more effective recognition methods [35][36][37][38] have been developed. For example, Gao et al [35] extended R1-PCA and proposed R1-2DPCA, which effectively protected the spatial structure of data and selected a nucleus norm as the distance metric to improve the classification effect in the classification stage.…”

“…Wang et al [37] proposed an Angle PCA model for image analysis, which considers the relationship between the variance of projection data and its reconstruction error, and used 2 l -norm as a robust metric. Bi et al [38] extended the 2 l -norm to a more flexible 2, p l -norm and proposed a generalized PCA model for image analysis, which can extract the important features of the image more accurately.…”

Underwater image processing and analysis: A review

“…The diversity and complexity of underwater environment has been induced widespread research interest lately. Xu et al [35] developed a novel approach for underwater target recognition in view of the generalized robust principal component analysis (GRPCA), which could extract the visual feature information from underwater images and was of great significance to the recognition and representation of subordinate images. Shen et al [51] explored a useful underwater target layered background framework on account of a frog eye visual information perception and processing scheme, which is able to separate salient objects from the image background with object contour.…”

Kapur’s Entropy for Underwater Multilevel Thresholding Image Segmentation Based on Whale Optimization Algorithm

“…In recent years, unmanned underwater vehicles (UUVs) with vision systems have been widely used to gather image information for analysis and research, in which underwater image segmentation is difficult to accomplish in machine vision. The quality of image segmentation directly affects the accuracy of target feature extraction and target detection [1]- [3]. The three-dimensional model of a UUV equipped with a vision system is given in Fig 1. Image segmentation is a fundamental and important technology in image processing and robotic vision, which is a key operation from image processing to image analysis, as well as one of the key target feature extraction, recognition and tracking.…”