DIGITAL IMAGE-BASED DENTAL CARIES CLASSIFICATION SYSTEM

Abstract

This invention relates to a digital image-based dental caries classification system that includes an image acquisition module for receiving a complete digital image of a tooth from a camera device, an image pre-processing module configured to convert a color image to a grayscale image, noise reduction using Gaussian filtering, and local contrast enhancement using Contrast Limited Adaptive Histogram Equalization (CLAHE), a feature extraction module configured to calculate the gradient orientation distribution over the entire image using the Histogram of Oriented Gradient (HOG) method by dividing the image into a number of cells and blocks for gradient normalization to produce a fixed-dimensional numeric feature vector without explicit identification of regions of interest, a classification module that includes a Multilayer Perceptron artificial neural network having one input layer,at least two hidden layers with non-linear activation functions, and one output layer that produces the probability of categories of carious teeth and non-carious teeth, where the network is trained using the Adam optimization algorithm and the categorical cross-entropy loss function, as well as a user interface module to display the classification results in the form of textual information.,