Detection of Defects in Dental with Support Vector

The process of dental defect analysis is to provide an efficient clinical support with less complexity in segmentation, better accuracy in foreground object detection and to provide local contrast and luminance invariant features using the various descriptors. The automatic decision support system includes adaptive threshold and support vector machine for dental disease prediction. The proposed segmentation approach, adaptive threshold is used to segment significant information from input image for extracting features. An adaptive threshold is determined based on pixel probability and variances. In feature extraction stage, segmented region will be used to extract features which describes about color and texture contents. The color features are extracted from HSV color space of selected region based on histogram analysis. Weber’s local descriptor (WLD) represents an image in terms of histogram features which are extracted from gradient orientation of an input. Discriminative Robust Local Ternary Pattern (DRLTP) based histogram features are used to differentiate the local object in terms of contrast, shape and illumination changes. Combined Color and texture features are used for SVM training with reference samples for its classification. SVM (Support Vector Machine) is the supervised learning model used to classify and identify the test dental image into either normal or abnormal defects based on supervised training with the radial basis kernel function. Finally the simulated result shows that utilized methodologies thereby providing better performance and good classification accuracy rather than prior methods and the identification of dental defects.