A Novel Color Feature for the Improvement of Pigment Spot Extraction in Iris Images

• Published in: Journal of Image and Graphics (United Kingdom)
• Main Author: Jabal M.F.A.; Abdullah A.N.H.; Najjar F.H.; Hamid S.; Khalid A.K.; Manan W.D.W.A.
• Format: Article
• Language: English
• Published: University of Portsmouth 2024
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85218261303&doi=10.18178%2fjoig.12.4.410-416&partnerID=40&md5=788a2df4062d47f76463790ffc17e45f

Feature extraction plays a vital role in the segmentation of regions of interest in medical images. While histograms offer a reliable method for analyzing color properties, the challenge of defining the pigment spot color has motivated the search for a practical feature for extraction. Consequently, analyzing the image using histograms and the HSV (Hue, Saturation, Value) color space led to the groundbreaking discovery of a reliable color feature and an exciting opportunity for pigment spot extraction. This study utilized 131 pigment spot images from the Miles Research datasets. The Region of Interest (ROI) was determined using a histogram color-based saturation intensity component, revealing new findings of thresholds ranging from 0.70 to 0.90. The results indicate that the proposed method achieved a Detection Rate (DR) of 37.1% (49 images), a False Acceptance Rate (FAR) of 14.5% (19 images), and a False Rejection Rate (FRR) of 48.4% (63 images). While the detection rate shows room for improvement, the proposed method significantly reduces the FAR to 14.5%, compared to 64.8% and 65.3% in color-based segmentation and simple color detection, respectively. This newfound feature contributes to improved accuracy and efficiency in medical image analysis, facilitating better patient diagnosis and treatment planning. © 2024 by the authors.