An improved edge detection algorithm using a modified discrete wavelet transform based on morphological thinner for noisy medical images
Shilpi Gupta, Ramesh Kumar Sunkaria · 2017
In Medical diagnoses, edges of any object in medical image are the conspicuous indication of the finding diseases in the human body. From this perspective, many edge detectors have been presented not only to detect edge, but also to connect neighborhood edge pixels in center edge points to preserve significant information. In this research paper, we present the method of improved edge detection, named as modified discrete wavelet transform with morphological thinner (DWT-T). In the method of DWT-T, multi-scale edge detector is used which detect edges in the presence of impulse noise at every resolution and morphological thinner, which connect all information contain edge points and disconnect undesired edge point. From the qualitative analysis, it has been found that proposed method exhibit more appropriate results as compared to previous edge detectors and gives adequate result in the area of medical image processing in terms of Peak signal to noise ratio (PSNR), Mean square error (MSE) and Normalized absolute error (NAE). A simulation result shows that proposed edge detector works very efficiently in the presence of Noise Density (up to 80%).