CBMIR algorithm appears to address key challenges in medical image retrieval, providing robustness to variations, efficient representation, and a high success rate in retrieval tasks. The emphasis on regional matching and query by example adds practical utility to the proposed method. The algorithm is designed to be robust to scaling and translation of objects within medical images. This robustness is crucial for handling variations in image sizes and positions. The attention is placed on an efficient representation and retrieval of medical images. The CBMIR method allows users to query by example, indicating a user-friendly and intuitive way to retrieve similar images based on a provided example. The algorithm’s efficiency and performance have been evaluated on a dataset of about 5000 simulated, realistic computed X-ray images. The dataset includes images selected from three large medical image databases, reflecting a diverse and comprehensive evaluation. The results of experiments indicate a success rate of more than 93%, which is deemed satisfactory. This suggests that the proposed algorithm performs well in retrieving relevant medical images from the dataset. The CBMIR algorithm appears to address key challenges in medical image retrieval, providing robustness to variations, efficient representation, and a high success rate in retrieval tasks.