A system for the automatic detection of pigment network

Pigment network is considered a key differential structure by dermatologists. Recently, different approaches have been proposed for the detection and characterization of this structure. This paper proposes an improved system for automatic detection of pigment network regions. The system starts by detecting the presence of pigment network using a bank of directional filters and a connected component analysis. After, a set of features, which characterize the pigment network's lines and the background, is extracted and used to validate the detected regions with AdaBoost. Finally, each lesion is classified regarding the presence or absence of pigment network The algorithm was tested in a dataset of dermoscopy images from Hospital Pedro Hispano (Matosinhos) achieving a SE = 78% and a SP = 77% for five fold cross validation.

[1]  G. Argenziano,et al.  Epiluminescence microscopy for the diagnosis of doubtful melanocytic skin lesions. Comparison of the ABCD rule of dermatoscopy and a new 7-point checklist based on pattern analysis. , 1998, Archives of dermatology.

[2]  Paul A. Viola,et al.  Robust Real-Time Face Detection , 2001, Proceedings Eighth IEEE International Conference on Computer Vision. ICCV 2001.

[3]  Murali Anantha,et al.  Detection of pigment network in dermatoscopy images using texture analysis. , 2004, Computerized medical imaging and graphics : the official journal of the Computerized Medical Imaging Society.

[4]  Rita Cucchiara,et al.  Line Detection and Texture Characterization of Network Patterns , 2006, 18th International Conference on Pattern Recognition (ICPR'06).

[5]  Paolo Carli,et al.  Epiluminescence microscopy of pigmented skin lesions , 2000 .

[6]  G. Fabbrocini,et al.  Dermoscopic image-analysis system: estimation of atypical pigment network and atypical vascular pattern , 2006, IEEE International Workshop on Medical Measurement and Applications, 2006. MeMea 2006..

[7]  S. Menzies,et al.  Frequency and morphologic characteristics of invasive melanomas lacking specific surface microscopic features. , 1996, Archives of dermatology.

[8]  M. Stella Atkins,et al.  A novel method for detection of pigment network in dermoscopic images using graphs , 2011, Comput. Medical Imaging Graph..

[9]  J. Mayer,et al.  Systematic review of the diagnostic accuracy of dermatoscopy in detecting malignant melanoma , 1997, The Medical journal of Australia.

[10]  Bill Triggs,et al.  Histograms of oriented gradients for human detection , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[11]  Jorge S. Marques,et al.  Detecting the pigment network in dermoscopy images: A directional approach , 2011, 2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[12]  Randy H. Moss,et al.  Automatic detection of blue-white veil and related structures in dermoscopy images , 2008, Comput. Medical Imaging Graph..

[13]  Jorge S. Marques,et al.  New Performance Evaluation Metrics for Object Detection Algorithms , 2004 .

[14]  Antonio Pietrosanto,et al.  Epiluminescence Image Processing for Melanocytic Skin Lesion Diagnosis Based on 7-Point Check-List: A Preliminary Discussion on Three Parameters , 2010 .