Extraction of the SLAM algorithms in gastric endoscopy

This research focuses on a step, in examining the structure of the stomach, which involving the collection of endoscopic images. This research aims to enable experts to make more accurate diagnoses. We explore techniques for improving these images and extracting features considering the challenges faced during stomach endoscopy. Our main objective is to enhance image quality while ensuring that vital information is preserved, and significant characteristics are quickly identified. To achieve this, we investigate strategies such as noise reduction methods like Gaussian blur. We also emphasize the importance of contrast enhancement through histogram equalization particularly using the Contrast Limited Adaptive Histogram Equalization (CLAHE) method, which improves contrast. Moreover, we employ image stabilization techniques that rely on flow-based methods and gyroscopic sensor data to address motion artifacts caused by camera or patient movement. These approaches ensure that our endoscopic images remain clear and precise throughout. As part of our study on feature identification techniques, we evaluate approaches including blob detection methods, edge detection algorithms. We offer an alternative called ORB (Oriented FAST and Rotated BRIEF), a more dependable method for extracting features. It is particularly well suited for real-time applications.