Global Pattern Classification in Dermoscopic Images

CONTENTS 6.1 Introduction ....................................................................................... 184

6.1.1 Global Patterns ...................................................................... 185 6.2 Global Pattern Detection ................................................................... 189

6.2.1 Texture Analysis for Global Pattern Classification by Tanaka et al. .......................................................................... 189 6.2.1.1 Segmentation ............................................................. 190 6.2.1.2 Feature Extraction..................................................... 190 6.2.1.3 Feature Selection ....................................................... 190 6.2.1.4 Classification.............................................................. 191

6.2.2 Parallel Pattern Detection by Iyatomi et al. .......................... 191 6.2.2.1 Segmentation ............................................................. 192 6.2.2.2 Feature Extraction..................................................... 192 6.2.2.3 Feature Selection ....................................................... 192 6.2.2.4 Classification.............................................................. 192

6.2.3 Color and Texture Analysis for Global Pattern Classification by Abbas et al. ...................................................................... 193 6.2.3.1 Segmentation ............................................................. 193 6.2.3.2 Feature Extraction..................................................... 193 6.2.3.3 Classification.............................................................. 194

6.2.4 Textons for Global Pattern Analysis by Sadeghi et al. .......... 194 6.2.4.1 Segmentation ............................................................. 195 6.2.4.2 Classification Method ................................................ 195

T&F Cat #K23910 — K23910 C006 — page 184 — 7/15/2015 — 21:20

6.2.5 Global Pattern Detection Based on Markov Random Fields by Serrano and Acha ............................................................. 196 6.2.5.1 Segmentation ............................................................. 196 6.2.5.2 Image Model and Feature Extraction ........................ 196 6.2.5.3 Classification.............................................................. 197

6.2.6 Model-Based Global Pattern Classification by Sa´ez et al. .............................................................................. 199 6.2.6.1 Segmentation ............................................................. 199 6.2.6.2 Feature Extraction..................................................... 199 6.2.6.3 Model-Based Classification ........................................ 200

6.3 Discussion........................................................................................... 204 References .................................................................................................. 206

6.1 INTRODUCTION Dermoscopy (also known as dermatoscopy, epiluminescence microscopy, incident light microscopy, and skin-surface microscopy) is a noninvasive technique to examine pigmented anatomic structures of the epidermis, dermoepidermal junction, and superficial papillary dermis that are not visible to the naked eye [1]. The abnormal structural features of melanoma can be identified, borderline lesions may be closely observed, and benign lesions can be confidently diagnosed without the need for biopsy using dermoscopy. Dermoscopy has been shown to be more accurate than clinical examination for the diagnosis of melanoma in a pigmented skin lesion [2, 3].