<![CDATA[Artificial intelligence models are becoming increasingly essential in biomedical research and healthcare services. Various healthcare organizations utilize information-based machine learning and image-processing methods for the diagnosis of diseases. This review delves explicitly into elucidating the challenges and considerations of developing unsupervised learning for clinical decision support systems in real-world contexts. In recent years, supervised and unsupervised deep learning have demonstrated promising medical imaging and image analysis outcomes. Unsupervised learning gathers data, draws insights from it, and makes data-driven judgments without bias, unlike supervised learning, which requires manual class labeling. A systematic review of unsupervised medical image analysis methods is presented here. This extensive review introduces diverse methodologies rooted in unsupervised classification for detecting diseases and analyzing images. Moreover, we offer insights into publicly available image benchmarks, datasets, and performance measurement details. Each method's strengths and weaknesses are thoroughly discussed, complemented by tabular summaries illuminating each category's outcomes. Additionally, the article furnishes detailed descriptions of the frameworks employed by each approach and the image datasets utilized.]]>
