Close range photogrammetry and self-organizing map for automatic diagnosing diseases

The output of close range photogrammetry is a finite set of 3D points which causes generating a discrete space. The dataset is a point cloud with no defined topology and structure. For diagnosing disease precisely using close range photogrammetry, specifically in an intelligent manner, it is necessary to reconstruct and parameterize a continuous and topology-definable surface from the measured points. Traditional methods for surface reconstruction such as interpolation and using approximation functions do not provide topological information and required details for shape analysis and disease diagnosis. According to SOM abilities for reconstruction of the space of measured points as a fully structured and topology-definable surface, in this research, a medical system has been designed and implemented by integration of SOM and close range photogrammetry. The research result shows that SOM is an effective tool for recognizing the pattern of measured points and generating a reference surface around affected areas. So, close range photogrammetry and SOM can be used to develop integrated systems as two complementary techniques for diagnosing diseases whose symptoms are visible or appear as deformations out of body and around the affected area.