Non-camera Wireless Capsule Endoscopy for Early Detection of Small Intestinal Tumors Utilizing Radio Signal Recognition

Early detection of small intestinal cancer is crucial due to subtle symptoms. This research addresses camera-based Wireless Capsule Endoscopy (WCE) limitations in visualizing specific Gastrointestinal (GI) regions: 1) visible tumors and 2) tumors outside WCE visibility. Our non-camera WCE employs radio signal recognition and Support Vector Machine (SVM) with a Radial Basis Function (RBF) kernel for straightforward classification of small intestinal tumor detection. Focused on S21 patterns in human voxel models, our simulation discerns dielectric variations in normal and tumorous tissues, replicating intricate characteristics. Initial results demonstrate efficacy in differentiating normal and tumor cases. This non-camera approach minimizes components, operates on low power for safety, promising widespread screening and diagnosis. Simulating S21 patterns using SVM with an RBF kernel enhances our system, potentially revolutionizing early cancer detection. With a $98.4 \%$ accuracy rate, our model excels in distinguishing normal from tumor tissues, elevating diagnostic precision. Nevertheless, despite the challenge of visually locating tumors, our non-camera WCE enables early tumor detection through cost-effective and rapid investigation, providing a preliminary screening before complex procedures. This research signifies a substantial stride in improving healthcare outcomes through cutting-edge technology.