Clustering is one of the most vital image segmentation techniques. However, proper image clustering has always been a challenging task due to blurred and vague areas near to concerned object boundaries. Therefore, rough set based clustering techniques like Rough k-means (RKM) has been employed in image clustering domain because rough set concept can handle the overlapping clusters to a great extent. RKM shows the performance in image clustering domain as a similarity based clustering model like K-Means and Fuzzy C-Means. Therefore, this paper presents an up-to-date review on rough set based image clustering approaches with their merits and demerits. The measures of similarity as well as the evaluation criteria for rough clustering are also discussed in this study. Beside that the key issues which are involved during the development of rough set based clustering models are investigated in this paper.

聚类是最重要的图像分割技术之一。然而，由于相关对象边界附近的模糊和模糊区域，适当的图像聚类一直是一项具有挑战性的任务。因此，基于粗糙集的聚类技术如 Rough k-means (RKM) 已被用于图像聚类领域，因为粗糙集的概念可以在很大程度上处理重叠的聚类。RKM 将图像聚类领域的性能展示为基于相似性的聚类模型，如 K-Means 和 Fuzzy C-Means。因此，本文对基于粗糙集的图像聚类方法及其优缺点进行了最新评论。本研究还讨论了相似性的度量以及粗聚类的评估标准。