Search

In this paper we extend the Shared Biological Image Manager (SBIM) system by incorporating a content-based image retrieval (CBIR) technique using machine learning concepts. For the classification it was applied a multiclass support vector machine (SVM) schema. Given a query image, the SVM approach performs its classification and the Euclidean distance is used to compute the similarity between the query and the images in the same class. The computational experiments show satisfactory classification accuracy and suitable retrieval results. Besides, the new SBIM was validated using integration tests which show that the methods implemented attend the proposed functionalities.

Panoramic x-ray images are very popular as a first tool for diagnosis in odontological protocols. Automating the process of analysis of such images is important in order to help dentist procedures. In this process, teeth segmentation of the radiographic images is an essential step. In this paper, we propose a segmentation approach based on a supervised learning technique for texture recognition. Firstly, feature extraction is performed by computing moments and statistical features. The obtained data are the input to a Bayesian classifier that, after training, can distinguish two classes of pixels: active (inside the target texture) or inactive (outside the teeth). In the experimental results we show that the methodology is a promising one for teeth segmentation in panoramic x-ray images and discuss its limitations.