Steel Surface Defect Classification Based on Discriminant Manifold Regularized Local Descriptor

Automated Visual Defect Detection for Flat Steel Surface: A Survey

“…As a classical operator, local binary pattern (LBP) is widely used to characterize local texture features of images, which has significant advantages of rotation and gray invariance. In 1994, LBP is first proposed by Ojala et al in [42], Later, LBP is frequently used to detect defects on flat steel surface [43][44][45]. In order to overcome the shortcomings of the original LBP (i.e., weak global descriptive and noise-sensitive), various LBP variants are developed based on changing the threshold or scale of the original LBP (refer to Fig.…”

Automated Visual Defect Classification for Flat Steel Surface: A Survey

“…LBP is one of the most successful local texture feature operators, which creates an intensity-and rotation-invariant binary descriptor and estimates the local contrast of an image based on the differences between adjacent pixels and central pixel, whose encoding mode and sampling rules are briefly given in Fig. 4, and it has been widely used to extract features of steel surfaces [11,[81][82][83][84][85][86][87]. In addition, some variants based on the original LBP have been proposed to overcome the limitations of LBP, such as noise sensitivity.…”

“…The earlier local descriptors, such as the LBP and HOG, subject to the hand-crafted definition and the limitations of the applications. In contrast, Zhao et al [87] proposed the discriminant manifold regularized local descriptor (DMRLD) based on the new viewpoint learning mechanism, which applies the manifold structure to regularize the local descriptor for describing the features of the image. Its core idea is to employ the learning strategy to establish the local information, while maintaining the original, discriminant, and intrinsic structure of the steel defect image.…”

“…Manifold learning [87] Be able to find the essence of from the observed phenomena, and find the internal law of data generation.…”

“…Luo et al [10,90] selected the NNC as the dissimilarity metric between two multi-region histograms to identify LBP-like histograms. Moreover, Zhao et al [87] simultaneously considered the variation and exemplar distances of the NNC to measure the similarity between the local models. The accuracy of this kind of conventional classifier primarily relies on the precision of the extracted features.…”