Can I read Eczema Skin Lesions Segmentation Using Deep Neural Network (U-Net) on EtoBox?

Eczema Skin Lesions Segmentation Using Deep Neural Network (U-Net) by Humaira Nisar; Ysin Ren Tan; Kim Ho Yeap is a book available to read on EtoBox.

What is Eczema Skin Lesions Segmentation Using Deep Neural Network (U-Net) about?

Atopic Eczema is a skin condition that causes dry, itchy, and inflamed skin. The treatment of eczema requires detection of the infected region and grading of eczema symptoms. Dermatologists assess eczema visually and use special forms to record their findings. This process is tiresome and uses a great deal of time. In addition, it introduces inter-observer and intra-observer variability in the results. Hence in this work a fully automated method is developed to segment eczema skin lesions. A five-stage U-Net is trained to perform segmentation on the eczema dataset that consists of 84 images of different severity levels. Four different color spaces i.e., the RGB color space, HSV color space, YCbCr color space and CIELAB color space are employed for the segmentation analysis. In addition, the effect of color space normalisation technique and Adaptive Light Compensation (ALC) for illumination compensation are also examined. For performance evaluation five metrics are used namely, accuracy, sensitivity, specificity, Positive Predictive Value (PPV) and Negative Predictive Value (NPV). The highest average segmentation accuracy of 87.44% is achieved by using 16 channels in the first stage

Author

Humaira Nisar; Ysin Ren Tan; Kim Ho Yeap

Publisher

Springer International Publishing AG

Published

2023

Language

EN

ISBN

9783031232398

Subjects

Engineering, Stem