The Research About Topic Extraction Method Based on the DTS-ILDA Model

Kernel-based Algorithms for Image Classification: A Review

“…The average local singularity is defined based on high-order statistics in a local sliding window, which is used to select the most informative nonlinear components for anomaly detection. A method based on the extracting block integrated features of KPCA was presented for extracting the characteristics of various ethnic groups in features of Chinese face images [3]. Firstly, KPCA has extracted the global features of the face image and the local feature of each subblock.…”

“…Kernel methods proved effective in the analysis of remote sensing images, by improving the results offered by traditional statistical methods, such as natural resource control, detection and monitoring of anthropic infrastructures, and agriculture inventorying [1]. Nowadays, kernel methods are standard techniques for many computer vision applications, including face recognition [2,3], bearing fault diagnosis [4] and online scene classification [5].…”

“…The second part addresses the classical and recent research in developing kernel methods for supervised image classification, semi-supervised image classification, and feature extraction (Sec. [3][4][5]. The third part devotes to the applications of kernel techniques in online learning, few-shot learning, deep learning and for image classification (Sec.…”

“…Then the typical multivariate analysis models could be generalized with the kernel trick to build their kernelized version: Kernel PCA, Kernel CCA, and Kernel OPLS [20][21][22]. These methods were used to diagnose bearing fault [4], and to extract features of face images of various ethnic groups in China [3] . For hyperspectral image classification, kernel nonparametric weighted feature extraction (KNWFE) [23] and kernel minimum noise fraction (KMNF) [24] have been widely used.…”