Systematic Development of Modified Hybrid Learning based Detection Scheme for White Blood Cancer Cells using Medical Image Processing Methodology with IoT Alert Mechanism
P Vinayagam, Nellore Vishnu Vardhan Reddy, Paturu Teja Krishna Kishore, Angelina Royappa
Abstract
The automated detection of white blood cell cancers like cancer and lymphoma is an active area of medicinal study that presents significant challenges. In this research, we unveil a brand-new, cutting-edge tool for aiding in the diagnosis of disorders affecting the white blood cells. In this work, we propose Hybrid Learning based Medical Image Evaluation (HLMIE), a novel deep learning based cancer cell classification technique. The proposed approach is grounded in traditional learning and classification techniques like Neural Network based Learning and Random Forest Classifier (RFC), and entails creating an automated system to aid doctors in making accurate diagnoses of the various forms of this illness. Acute Myolegenous Leukemia (AML) and Acute-Lymphoblastic-Leukemia (ALL) are the two categories, and they share several symptoms that might make diagnosis difficult. Several techniques involving machine learning and deep learning have been employed in the past to try to make predictions about blood cancer; however these researches have their flaws. Therefore, this research proposes a deep learning/classification logic/medical image processing hybrid model to enhance prediction accuracy. Different degrees of prediction, analysis, and learning methods are incorporated into the suggested hybrid learning model enabled with image processing, and various learning criteria, such as development of learning and assessment precision with regard to epochs, are utilized. In addition, limiting fine-tuning of the system’s output to actual experts and this scheme provides better accuracy in results with respect to better classification performance with prediction logics using hybrid learning methodology, in which the proposed scheme attains 97.84% of prediction accuracy and the reduce the image noise ratio as 2.36% as compared with the conventional classification and learning algorithms. The resulting section shows the proper proof of this quote with exact histographical specification. The resulting summary of the disease prediction is reported properly to the respective care taker or an individual with respect to the principles of trending technology called Internet of Things (IoT).