Bi-dimensional Empirical Mode Decomposition and Nonconvex Penalty Minimization L q (q = 0.5) Regular Sparse Representation-based Classification for Image Recognition
- 24 Downloads
This paper reports an innovative pattern recognition technique for fracture microstructure images based on Bi-dimensional empirical mode decomposition (BEMD) and nonconvex penalty minimization L q (q = 0.5) regular sparse representation-based classification (NPMLq-SRC) algorithm. The detailed procedures of this work can be divided into three steps, i.e., the preprocessing stage, the feature extraction stage and the image classification stage. We test and validate the proposed method through real data from metallic alloy fracture images. The case verification results show that our proposal can obtain a much higher recognition accuracy than the conventional Back Propagation Neural Networks (BPNN for short), the L1-norm minimization sparse representation-based classification (L1-SRC) and the BEMD combined with L1-norm minimization sparse representation-based classification (BEMD+L1-SRC) methods, respectively. Specifically, the proposed BEMD+NPMLq-SRC (q = 0.5) method outperforms the BEMD+L1-SRC method by 3.33% improvement of the average recognition accuracy, and outperforms L1-SRC method by 14.06% improvement of the average recognition accuracy, respectively.
Keywordsimages recognition fracture microstructure images (FMI) BEMD method feature extraction NPMLq-SRC compressed sensing
Unable to display preview. Download preview PDF.
- 4.K. Komai, K. Minoshima, and S. Ishji, “Recognition of different fracture surface morphologies using computer image processing technique,” Int. J. Ser. A, Mech. Mater. Eng. 36 (2), 220–227 (1993).Google Scholar
- 5.K. Minoshima, T. Nagasaki, and K. Komai, “Automatic classification of fracture surface morphology using computer image processing technique,” Jpn. Soc. Mech. Eng. 56 (625), 1319–1323 (1990).Google Scholar
- 6.Z. N. Li, Y. Sun, J. W. Yan, S. R. Long, and Y. C. Yang, “Study on the recognition method of metal fracture images based on Grouplet-RVM,” Chn. J. Sci. Instrum. 35 (6), 1347–1353 (2014).Google Scholar
- 8.Y. H. Yan, J. H. Gao, Y. Liu, Y. G. Cao, and S. C. Lei, “Recognition and classification of metal fracture surface models based on wavelet transform,” Acta Metall. Sin. 38 (2), 309–314 (2002).Google Scholar