Pattern coloring By D2NN and FLSA-SVM based on Probabilistic ML model for diabetic patient

• Published in: Multimedia tools and applications 2024-01, Vol.83 (21), p.60689-60716
• Main Authors: Shabdiz, Marzieh, Azarbar, Ali, Azgomi, Hossein
• Format: Article
• Language: English
• Subjects: Algorithms; Color; Computer Communication Networks; Computer Science; Data points; Data Structures and Information Theory; Diabetes; Diffraction patterns; Multimedia Information Systems; Redundancy; Signs and symptoms; Special Purpose and Application-Based Systems; Statistical analysis; Support vector machines; Track 2: Medical Applications of Multimedia; Uniqueness
• ISSN: 1573-7721; 1380-7501; 1573-7721
• DOI: 10.1007/s11042-023-17199-4

Iridology can detect a particular disease and activity of specific body areas; any disease depends on the color of the background of the IRIS. The type of spots and their position with unique colors is necessary. This paper represented blue IRIS with brown spots and gray circles for the intensity of glucose in diabetic patients. This study provided a pattern for probabilistic colored IRIS. We calculated the variant IRIS backgrounds with colored spots in the foreground and trained in the color space and iteration points by the l*a*b method; It also clustered the dots using FLSA-SVM in a colored diabetic IRIS. However, it had iterations in the scan spiral curve; every data point could redundant. This solution eliminated data redundancy with determined data points (radius) and seven zones in the spiral curves. It limited the vectors and dots by taking probability rules followed by color symptoms, subsequently. This method used the FLSA-SVM algorithm to cluster points in groups near vectors of the IRIS. The algorithm detected signs in different colors from the background of the IRIS accurately, and selected a diabetic model for complex detection and similarity with other diseases, such as liver and pancreatic cancer. This algorithm labeled the data that they had symptoms for a more detailed design. In addition, it will combined colors to create a new pattern in the target function. This study proposed a constraint argument on the multi-color model, and added similarities of colors in the new model with area-weighted to Error estimation. The labels used the k-means method to analyze unique decision boundaries in diffraction-colored edges. This study generated a spiral pattern to detect the layer of the iris and their spots with the colored method by two-spiral static values.