Volume 9, Issue 3 (lasers in medicine 2012)
lmj 2012, 9(3): 6-14 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Improvement of Erythema parameter auto Measurment algorithm of The PASI standard in psoriasis disease. lmj 2012; 9 (3) :6-14
URL: http://icml.ir/article-1-261-en.html
URL: http://icml.ir/article-1-261-en.html
Abstract: (12887 Views)
Background and Objectives: Psoriasis is an auto immune skin disease. In this disease, immunity system is disordered and skin is covered by red plaques. The plaques may be covered by silvery-scales. This disease has no definite cure and can only be controlled by drugs. The PASI standard is employed to evaluate effectiveness of drugs on patient. It consists of four parameters: area, erythema (redness), scaliness and thickness of plaques. Physicians visually assess the erythema parameter of PASI which comes to different results if performed by different physicians. We aim to improve the erythema parameter measurement algorithm in this research.
Material and Methods: The target is to design a pattern detection system (classifying redness of plaque) based on 56 lesions of disease. In this work, mean statistical moment descriptors of HSV color space components of skin and plaques are used which are correlated with physician diagnosis. When physician opinions disagree, a right opinion is selected by the proposed disagreement removal algorithm. Consequently, for producing and evaluation of a trustful model a random data appointment method is used. Finally, Digital image set of lesions is statistically analyzed and a regression model including three variables is proposed for optimum decision function to classify plaques.
Results: 1) improved accuracy with maximum unit error 2) improved prediction with 78 percent correlation, and 3) lower number of independent variables compared to the previous methods
Conclusion: In this work, proposed disagreement removal algorithm of evaluator and random data appointment method improve Erythema parameter auto measurement compared to previous algorithms. Accuracy of predicting Erythema score indicates higher performance of the proposed method compared to the previous works.
Material and Methods: The target is to design a pattern detection system (classifying redness of plaque) based on 56 lesions of disease. In this work, mean statistical moment descriptors of HSV color space components of skin and plaques are used which are correlated with physician diagnosis. When physician opinions disagree, a right opinion is selected by the proposed disagreement removal algorithm. Consequently, for producing and evaluation of a trustful model a random data appointment method is used. Finally, Digital image set of lesions is statistically analyzed and a regression model including three variables is proposed for optimum decision function to classify plaques.
Results: 1) improved accuracy with maximum unit error 2) improved prediction with 78 percent correlation, and 3) lower number of independent variables compared to the previous methods
Conclusion: In this work, proposed disagreement removal algorithm of evaluator and random data appointment method improve Erythema parameter auto measurement compared to previous algorithms. Accuracy of predicting Erythema score indicates higher performance of the proposed method compared to the previous works.
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
