Diamond clarity refers to the absence of tiny, natural inclusions (imperfections) inside a diamond or on its surface. Almost all diamonds contain their own unique inclusions due to their natural formation process. In this paper, a new inclusion extraction approach is developed to accurately separate the regions of interest in a diamond clarity image and then identify the image features of each region. The inclusion regions can be successfully distinguished from other types of signals. The findings of the theoretical optical analysis facilitate the image processing development and also reduce its complexity and operation time. The experimental results verify the effectiveness and robustness of the proposed inclusion extraction approach. The diamond inclusions can be accurately extracted from the captured diamond clarity image. The extracted inclusions can also be converted to their actual size as seen by the naked human eye. The proposed approach is verified to be significantly less sensitive to noise than existing approaches and unaffected by the fluctuations in illumination.