A schemefor patient study retrieval from 3D brain MR volumes

The paper presents a pipeline for case retrieval in magnetic resonance (MR) brain volumes acquired from biomedical image sensors. The framework proposed in this paper, inputs a patient study consisting of MR brain image slices and outputs similar patient case studies present in the brain MR volume database. Query slice pertains to a new case and the output slices belong to the previous case histories stored in the database. The framework could be of immense help to the medical practitioners. It might prove to be a useful diagnostic aid for the medical expert and also serve as a teaching aid for students and researchers in the medical field. Apart from diagnosis, radiologists can use the tumor location to past case studies relevant to the present patient study, which can aid in the treatment of the patients. Similarity distance employed in this work is the three dimensional Hausdorff distance which is significant as it takes into account the spatial location of the tumors. The preliminary results are encouraging and therefore the scheme could be adapted to various modalities and pathologies.