Original Article
Author Details :
Volume : 10, Issue : 4, Year : 2024
Article Page : 216-225
https://doi.org/10.18231/j.ijn.2024.045
Abstract
Background: Though and even though advancements are high during last two decades, yet ‘breast—cancer’ continues to be a sizable causal transience amongst females (womankind) globally. Whilst mammography has distinctly lowered death rates, detection and categorization and accuracy of breast sizes (forms in quantities) in mammograms remains argue. Thus we present a new method of detection and discovery of key region-of-interest(ROI) plus its labeling as natural usual ‘0’, benign ‘1’, or cancerous ‘2’ sizes.
Aim: For sorting of identified ‘ROIs’ as usual, benign or malicious exactly.
Materials and Methods : We use the removal of temporally sequential digital mammograms, merged via machine-learning (ML). We estimated feature-selection algorithmic techniques (ATs) over the novel database (data—set) comprises images~circa352 as of 88 diseased exactly noted by sites-of-mass. 98 feature extractions were done and classified by 8 dissimilar feature-selection ATs to detect the best discriminative features. We then computed 10classifiers via first in first out (FIFO) diseased plus a k-fold cross validation. In our computation, we observed the artificial neural nets (ANNs) appeared as the highest classifier with 99.5?curacy, 0.98AUC (for label ‘0’), 0.92AUC (for label ‘1’, plus 0.94AUC (for class ‘2’).
Results: Our findings showed significant progress in divergence (disparity) through existing advanced high-tech algorithms and techniques, which feature worth of using temporally successive mammographs in amalgamation, with advanced computational intelligence ML, aimed at the exact sorting of identified ‘ROIs’ as usual, benign or malicious.
Conclusion: Clinical significancy/ when clinically applied, this study improves the clinical prognosis precision of breast-cancer possibly steering to reliable diseased outcomes plus added personalized therapeutic attempts.
Keywords: Breastcancer (BC), Computer Aided Diagnosis (CAD), Digital mammograph, Machine Learning (ML), Prediction, Region of Interest (RoI), Temporal subtraction, Machine learning
How to cite : Devi G N R, Raju V R, Improving breast-cancer diagnostics: A study using AI-ML with computational prognostics and feature extraction. IP Indian J Neurosci 2024;10(4):216-225
This is an Open Access (OA) journal, and articles are distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 License, which allows others to remix, tweak, and build upon the work non-commercially, as long as appropriate credit is given and the new creations are licensed under the identical terms.
Received : 05-10-2024
Accepted : 26-10-2024
Viewed: 153
PDF Downloaded: 25