Enhancing Wireless Capsule Endoscopy Images for Illumination and Noise

Wireless Capsule Endoscopy (WCE) is a noninvasive diagnostic procedure to inspect the gastrointestinal tract. The images generated by WCEs are accompanied by noise and plagued from low brightness and contrast. Existing methods to enhance brightness, contrast, and reduce noise in WCE images often compromise crucial image details and textures, affecting the accuracy of diagnoses. This study has aimed at enhancing the brightness and contrast of images and reducing noise while preserving their details. In our proposed approach, the images are initially enhanced by Gamma correction then images are divided into smooth and non-smooth areas according to the texture information. Illumination and reflectance components of each area are then extracted from the image using a Homomorphic filter. Finally, an anisotropic diffusion operation is applied to the output of the Homomorphic filter, and the appropriate recursive coefficient is selected automatically for areas with different textures to remove noise. Quantitative analyses indicate that our proposed method enhances brightness and contrast while maintaining a more natural appearance and better preserving edges compared to existing methods. Furthermore, these analyses demonstrate that our method not only removes noise more effectively but also better preserves the structure and details of the images. Lastly, qualitative comparison visually demonstrated that the proposed method preserves more details, texture, and content than its counterparts.