Robotic Endoscopic for Smart Diagnosis

Developing healthcare systems for disease diagnosis and treatment require medical image segmentation as a vital prerequisite. U-Net, which is a U-shaped architecture, has become the standard in the segmentation of intestine images and has achieved remarkable success. The proposed approach is a novel medical image segmentation method that uses the TransUNet architecture, which combines both Transformers and U-Net, as an effective alternative. The system presents a review of the latest developments in robotic endoscopic systems, including their design, control algorithms, and clinical applications. The Transformer encodes tokenized image patches from a Convolutional Neural Network (CNN) feature map to extract global contexts, while the decoder up samples the encoded features and combines them with high-resolution CNN feature maps to allow precise localization. The Transformers serve as robust encoders for medical image segmentation tasks, while the combination with U-Net improves the recovery of localized spatial information and enhances finer details