3D Content-Based Retrieval for T1 Weighted Contrast Enhanced Magnetic Resonance Brain Database using Multiple Features

Advancement in medical imaging modality and the fast development of the internet is generating a larger database of high-resolution 3D Volumes of medical data. In the medical field, the Computer-aided systems that support disease diagnosis and correlation of different subjects with similar cases are of great application. Retrieval of 3D medical data with high resolution from a very large database, made the feature extraction of voxels, a vivid area of research. The paper exploits the efficiency of 3D wavelet transform and 3D Run-length matrices for representing volumetric texture feature. The system proposed an efficient 3D MRI scan retrieval from a 3D database, similar to the provided query. As for the retrieval precision, the proposed 3D CBR system achieves a precision rate of 90.12%. This paper evaluates the proposed CBIR system’s performance, for the Euclidean and the Manhattan distance metric.