Endoscopic Image Processing Method for Virtual Chromoendoscopy with Controlled Color Transform and Local Contrast Accuracy

In recent years, the development of medical video systems is accompanied by continuous improvement of visualization techniques, which promotes more accurate diagnosis of pathological changes in internal organs and tissues of patients. One of such technologies is virtual chromoendoscopy, which helps to improve the diagnostic value of endoscopic images. Known implementations in modern hardware are I-SCAN and FICE. In addition, there are works devoted to image processing algorithms simulating narrow-spectral imaging. This study introduces a novel approach to virtual chromoendoscopy (VCE) that utilizes digital image processing techniques applied to images captured under white light. A key characteristic of this method is its implementation of local nonlinear processing for each color channel. The effectiveness of this approach was evaluated using the KVASIR open database of endoscopic images, demonstrating significant improvements in color contrast. This innovative technique enables software-based image processing to achieve visualizations that align with cutting-edge advancements in VCE.