| Summary: | RGB colour variegations are useful features used by the domain's experts in their morphological learning method for skin disease classification. With the advancement of the computer vision technology, not only these features can be quantified in the digital image restoration and enhancement but also can be used as input parameters of an intelligent diagnostic system. In this report, several psoriasis lesion group are been studied for grayscale color features extraction. The experimental work involved clinical guttate lesion images where they are processed to produce the average Gaussian mean and standard deviation indices using the conventional algorithm. Normal and differential quantified indices gained under controlled environment are then mapped with another set of images from the same and other groups of the psoriasis lesion. The grayscale clustering plots together with each scale index distance from the reference indices are observed and analyzed. Finally, inference statistical tests are applied to conclude the findings. Outcome of the results show only guttate and erythroderma are distinguishable in grayscale mode.
|
|---|