Advances in syntactic imaging techniques for perception of medical images

Abstract

With the advances in pattern recognition technology, information systems have been applied to medical databases throughout the world. The most important problems in such access are how the pathological lesions can be efficiently recognized and how information in the respective databases can be summarized to enable intelligent selection of interesting cases of illness based on visual queries. The main goal of this paper is a presentation of the possibilities of the application of structural graph grammar-based methods of computer medical image analysis to the recognition and creation of a semantic description of the symptoms of lesions in the forms of local stenoses of the lumen of coronary arteries, and to the detection of pathological signs in upper parts of the ureter ducts and renal calyxes. An analysis of the correct morphology of these structures is possible owing to the application of sequence and graph methods from the group of syntactic methods of pattern recognition. In the analysis of coronary artery images, the main objective is computer-aided early diagnosis of the different forms of cardiac ischaemic diseases. In the analysis of kidney radiograms the main goal is the recognition of local irregularities in ureter lumens, and examination of the morphology of the renal pelvis and calyxes. The approach presented also allows the problem of image semantic indexing to be solved for feature-based retrieval in a specialist medical database.

Keywords:

• image analysis
• medical imaging
• syntactic pattern recognition
• artificial intelligence
• image understanding
• computer-aided diagnosis
• feature extraction