CHRONOLOGICAL OPTIMIZATION-BASED DEEP NEURO-FUZZY NETWORK FOR COLON CANCER DETECTION IN IOT

Colon cancer is one of the leading causes of cancer-related deaths worldwide, with early detection being pivotal for improving survival rates. Integrating the Internet of Things (IoT) with advanced computational methods allows real-time and automated diagnosis, reducing reliance on manual assessments and accelerating decision-making. The accurate analysis using the conventional diagnostic methods like Wireless Capsule Endoscopy (WCE) remains a significant challenge due to the complexity of identifying cancerous regions. This paper introduces a novel Chronological Student Psychology Optimization Algorithm-based Deep Neuro-Fuzzy Network (CSPOA-based DNFN) framework for efficient and accurate colon cancer recognition. The patient images are gathered through IoT, and data routing is optimized at the Base Station (BS) using the Multi-Objective Fractional Gravitational Search Algorithm (MOFGSA). Preprocessing is performed using a median filter and then, the segmentation is performed by using the SegNet. Thereafter, features like Local Gradient Pattern (LGP), Convolution Neural Network (CNN), and statistical features are extracted. The DNFN model is used to detect colon cancer, which is trained by CSPOA. The CSPOA is the combination of chronological concepts and the Student Psychology-based Optimization Algorithm (SPBOA). Also, the developed approach achieved the highest testing accuracy, sensitivity, specificity, precision, and [Formula: see text]-measure of 88.7%, 90.5%, 87.5%, 90.8%, and 88.8%, respectively. Due to these better results, this work facilitates early-stage detection of colon cancer, which is crucial in reducing mortality rates.