TY - JOUR
T1 - Psoriasis image representation using patch-based dictionary learning for erythema severity scoring
AU - George, Yasmeen
AU - Aldeen, Mohammad
AU - Garnavi, Rahil
N1 - Funding Information:
The research is supported by the University of Melbourne , Department of Dermatology at the Royal Melbourne Hospital, Victoria, Australia and Pfizer Australia Pty Ltd , IIR tracking number WI173634 . Authors also would like to thank University of Melbourne for allowing us to use Spartan HPC cluster for performing our extensive experiments. Yasmeen George received her B.Sc. and M.Sc. degrees in computer science from Faculty of Computer and Information Sciences, Ain Shams University, in 2008 and 2013, respectively. Currently, she is studying her PhD in Department of Electrical and Electronic Engineering, University of Melbourne. Her research interests include image processing, medical image processing, computer vision, machine learning, and artificial intelligence. Mohammad Aldeen obtained his BAE degree from Baghdad University, Iraq, MES degree from the University of Michigan, Ann Arbor, and PhD from Brunel University London, UK. Between 1982 and 1983, he was a postdoctoral research fellow at the Department of Electrical Engineering, Brunel University London, UK. He is currently an Associate Professor and director of Future Grid Research Centre. His research interests include medical image processing, and modelling and control of complex systems. Rahil Garnavi is a senior researcher at IBM Research Australia and an honorary research fellow at University of Melbourne. She received her BSc from Amirkabir University of Technology, MSc from the University of Isfahan and PhD from the University of Melbourne. She joined IBM in 2011 as a research scientist. She has published in numerous journals and conference proceedings and holds several patents in the areas of medical image analytics, machine learning, and artificial intelligence.
Publisher Copyright:
© 2018
PY - 2018/6
Y1 - 2018/6
N2 - Psoriasis is a chronic skin disease which can be life-threatening. Accurate severity scoring helps dermatologists to decide on the treatment. In this paper, we present a semi-supervised computer-aided system for automatic erythema severity scoring in psoriasis images. Firstly, the unsupervised stage includes a novel image representation method. We construct a dictionary, which is then used in the sparse representation for local feature extraction. To acquire the final image representation vector, an aggregation method is exploited over the local features. Secondly, the supervised phase is where various multi-class machine learning (ML) classifiers are trained for erythema severity scoring. Finally, we compare the proposed system with two popular unsupervised feature extractor methods, namely: bag of visual words model (BoVWs) and AlexNet pretrained model. Root mean square error (RMSE) and F1 score are used as performance measures for the learned dictionaries and the trained ML models, respectively. A psoriasis image set consisting of 676 images, is used in this study. Experimental results demonstrate that the use of the proposed procedure can provide a setup where erythema scoring is accurate and consistent. Also, it is revealed that dictionaries with large number of atoms and small patch sizes yield the best representative erythema severity features. Further, random forest (RF) outperforms other classifiers with F1 score 0.71, followed by support vector machine (SVM) and boosting with 0.66 and 0.64 scores, respectively. Furthermore, the conducted comparative studies confirm the effectiveness of the proposed approach with improvement of 9% and 12% over BoVWs and AlexNet based features, respectively.
AB - Psoriasis is a chronic skin disease which can be life-threatening. Accurate severity scoring helps dermatologists to decide on the treatment. In this paper, we present a semi-supervised computer-aided system for automatic erythema severity scoring in psoriasis images. Firstly, the unsupervised stage includes a novel image representation method. We construct a dictionary, which is then used in the sparse representation for local feature extraction. To acquire the final image representation vector, an aggregation method is exploited over the local features. Secondly, the supervised phase is where various multi-class machine learning (ML) classifiers are trained for erythema severity scoring. Finally, we compare the proposed system with two popular unsupervised feature extractor methods, namely: bag of visual words model (BoVWs) and AlexNet pretrained model. Root mean square error (RMSE) and F1 score are used as performance measures for the learned dictionaries and the trained ML models, respectively. A psoriasis image set consisting of 676 images, is used in this study. Experimental results demonstrate that the use of the proposed procedure can provide a setup where erythema scoring is accurate and consistent. Also, it is revealed that dictionaries with large number of atoms and small patch sizes yield the best representative erythema severity features. Further, random forest (RF) outperforms other classifiers with F1 score 0.71, followed by support vector machine (SVM) and boosting with 0.66 and 0.64 scores, respectively. Furthermore, the conducted comparative studies confirm the effectiveness of the proposed approach with improvement of 9% and 12% over BoVWs and AlexNet based features, respectively.
KW - Computer-aided system
KW - Multi-class classifier
KW - Patch-based feature extraction
KW - Psoriasis erythema severity scoring
KW - Sparse representation
KW - Unsupervised dictionary learning
UR - http://www.scopus.com/inward/record.url?scp=85042909278&partnerID=8YFLogxK
U2 - 10.1016/j.compmedimag.2018.02.004
DO - 10.1016/j.compmedimag.2018.02.004
M3 - Article
C2 - 29524784
AN - SCOPUS:85042909278
SN - 0895-6111
VL - 66
SP - 44
EP - 55
JO - Computerized Medical Imaging and Graphics
JF - Computerized Medical Imaging and Graphics
ER -