AIMMS system framework: Automatic dental pathologies recognition from DICOM files

This paper is made within the context of the research UEFISCDI project no. 31/2014, “AIMMS - Application for Using Image Data Mining and 3D Modeling in Dental Screening”. The project is interdisciplinary and wants to build a system framework composed of several sub-systems: dental pathology recognition part, 3D printing part, user interface part. The main emphasis of this article regards the pathology recognition part that is one of the core areas of our framework. Dental pathologies can be detected automatically from Cone Beam Computed Tomography (CBCT) data extracted from DICOM files. Dental CBCTs have a lower dose of radiation compared to Computerized Tomography (CT) scans, offering 3D information that sustains the diagnosis, treatment planning and analysis of the patient oral cavity. Our approach consists of novel procedures like semantically annotating the tomography which is facile to be processed. The analyzed oral pathologies are edentation and dental cavities using our own adaptive threshold filter and edge detection algorithm. An important role is played by the knowledge base of the mouth cavity during the overall processing of the DICOM input data. The cost of our solution is lower compared to the existent systems on the market, incorporating features that sustain its synergistic property.