Fusion of the Color Channels Using Densitometry for an Acute Lymphoblastic Leukemia Detection
Muhammad Yusuf, Arif Muntasa, Fitri Damayanti · 2019
Leukemia, especially Acute Lymphoblastic Leukemia (ALL) is a dangerous disease which attacks the bone marrow. One of the challenges of detecting the ALL quickly is the limitation of the methods to segment and find the main features. This paper aims to develop a model for the ALL detection using the color Densitometry features selection. Our proposed method purposes to improve the green channel original image through the contrast limited adaptive Histogram equalization and extract the main features based on the Densitometry. However, we utilized all of the image channels to find the main features. i.e., the red energy, the entropy, the red shanon entropy, the red log entropy, the green energy, the green entropy, the green shanon entropy, the green log entropy, the blue energy, the blue entropy, the blue shanon entropy, and the blue log entropy. We employed the acute lymphoblastic leukemia image database 2 (ALL-IDB2) to assess the performance of our proposed method. The proposed method has produced a good accuracy at 93.08% and standard deviation at 0.020. The novelties of this research are a new model to improve an image through the green channel, enhance the green channel images using the contrast limited adaptive Histogram equalization (CLAHE) method, and utilize unsharp mask filtering to obtain the best image segmentation. Furthermore, this paper contributes by developing a new model to detect the ALL using color Densitometry for other researchers in computer vision development.