An Image Fusion Method Based on Multi-angle Illumination Targeted for Medical Endoscopes

• Main Authors: Wen-Chi Huang, 黃文圻
• Other Authors: Yuan-Chen Liu
• Format: Others
• Language: en_US
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/26139548019066575550

碩士 === 國立臺北教育大學 === 資訊科學系碩士班 === 102 === In this thesis, a novel method for improving the image quality of the endoscopy for the gastrointestinal (GI) tract is presented. The proposed method takes pictures with multi-angel illuminations one by one, and then combines these pictures to form one image without specular highlights and shade. During a GI tract inspection, the probability of finding illnesses by doctor decreases without using clear pictures taken from endoscopes. During a typical inspection with the electronic endoscope, the light source machine is set at the average light sensing mode to automatically control the aperture. In this mode, near objects of pictures will show many specular highlights, and this makes endoscopists hard to determine what is on the surface. However, the peak mode shows better quality with near objects, but the far objects will be shadowed and hard to be distinguished. To solve these problems, the proposed method is that taking several pictures by an endoscope with different angles illumination, and the brightness and intrinsic colors from these images are calculated. Finally, the proposed method uses the results to merge the images as one with more details and correct colors. The high dynamic range image is also implemented for comparing with the image generated by the proposed method, and the results will be evaluated by several doctors to prove the feasibility of the proposed method.