Open AccessA Sub-band Anisotropic Diffusion Technique for de-speckling of Ultrasound Images of Breast Cancer
Author(s) -
Mayank Singh,
Indu Saini,
Neetu Sood,
Jasleen Saini
Publication year - 2021
Publication title -
aijr proceedings
Language(s) - English
Resource type - Conference proceedings
ISSN - 2582-3922
DOI - 10.21467/proceedings.114.10
Subject(s) - thresholding , artificial intelligence , wavelet , computer science , breast cancer , feature extraction , anisotropic diffusion , feature (linguistics) , pattern recognition (psychology) , ultrasound , breast ultrasound , noise (video) , mammography , wavelet transform , computer vision , signal to noise ratio (imaging) , image (mathematics) , cancer , medicine , radiology , linguistics , philosophy , telecommunications
Ultrasound imaging technique finds crucial application in clinical diagnosis of breast cancer. Presence of noise in ultrasound image due to different factor degrades the image quality and so the accuracy of diagnosis. Wavelet thresholding have been used from very beginning for de-noising of ultrasound image. Here in this paper we propose an intervention of anisotropic diffusion techniques in wavelet thresholding. In wavelet thresholding the thresholding operation usually applied after various feature extraction step, but in this study, we proposed to use a combinational approach. The approach reduces computational complexity of previous techniques. The proposed technique provides a Peak Signal to Noise Ratio of 28.46 and Mean Square Error of about 92.5537. The technique was practiced over large dataset of breast cancer images.