Performance Analysis of Panoramic Dental X-Ray Images Using Discrete Wavelet Transform and Unbiased Risk Estimation

Abstract

The images acquired in any modality will tend to have imposed noise in it. Denoising has to be done as a major step to perform further analysis. The denoising technique employed should be capable of removing all the noises present in the image. The dental panoramic X-ray images are denoised using discrete wavelet transform (DWT) and orthonormal wavelet transform (OWT) with Stein’s unbiased risk estimator (SURE). The performance analysis is done based on mean squared error (MSE) and peak signal-to-noise ratio (PSNR) values obtained from both the techniques. The denoising process preserves the information present in the signal like edges of the diseased signal. The denoised image is subjected to feature extraction and thresholding process to extract the diseased portion of the teeth. Artificial neural network is built to classify the images as normal or diseased teeth based on the extracted features. The performance analysis shows that orthonormal wavelet transform with Stein’s unbiased risk estimator performs well with good PSNR and MSE value.