Lossless Compression Based Optimization of DICOM Images Using Machine Learning

Modern medical knowledge and practice rely heavily on medical imaging. DICOM (Digital Imaging and Communications in Medicine) is the standard format used to store medical images after a patient's scan. However, DICOM files are large and require specialized viewers, which are not always available. This can be problematic when doctors need to share images quickly for treatment discussions. To address this, medical images can be converted to more accessible formats like JPG or PNG, which require less storage and are easily shareable and readable on various devices like smartphones, tablets, and computers. This paper explores the use of machine learning (ML) algorithms to optimize the compression ratio of medical images while maintaining high quality. Specifically, the Radial Basis Function Neural Network (RBFNN) is trained to predict the ideal compression ratio for CT scan images. Experimental results show that RBFNN outperforms other methods, achieving a 94 % recognition rate in the first scenario. In a second scenario, it achieves a 90 % accuracy with the optimal compression deviation set to 1, highlighting the effectiveness of ML for medical image compression.