Papers by Keyword: OCT Image

Paper TitlePage

Authors: Jing Liu, Xiao Lin Tian, Yan Kui Sun
Abstract: The traditional total variational (TV) model performs well for most noise image. However, the method will lose some information and details for the image which has rich texture and tiny boundary. Therefore, according to the requirements of the OCT pearl image, a novel denoising approach based on the TV model is proposed in this paper. This method combined the adaptive image denoising model and the novel fidelity term. Numerical experiments show that the proposed method can remove the noise while preserving significant image details. At pearl OCT image the method achieves at least 0.1dB gain over other existing denoising methods for Signal-Noise Ratio (SNR) measurement and Peak Signal-Noise Ratio (PSNR) measurement.
1383
Authors: Si Wei Huang, Ang Zhang, Xiao Lin Tian, Yan Kui Sun
Abstract: An edge detection algorithm which is applied to anterior chamber OCT images has been proposed. The algorithm firstly uses multi-structure elements to detect edge on gray level value differences on the same scale, and introduces dynamic adaptive weight to make re-fusion of pixels to gain a multi-structure element morphological edge detection image on the same scale, then confirms weight value and makes multi-scale fusion according to the noise immunity of different scale structure elements to gain the final edge detection image. The simulated results have obvious edge features,it can effectively avoid the occurrence of mutational pixels on the OCT image edge results, compared to traditional edge detection algorithms.
70
Authors: Si Wei Huang, Ang Zhang, Xiao Lin Tian, Yan Kui Sun
Abstract: An edge detection algorithm which is applied to anterior chamber OCT images has been proposed. The algorithm firstly uses multi-structure elements to detect edge on gray level value differences on the same scale, and introduces dynamic adaptive weight to make re-fusion of pixels to gain a multi-structure element morphological edge detection image on the same scale, then confirms weight value and makes multi-scale fusion according to the noise immunity of different scale structure elements to gain the final edge detection image. The simulated results have obvious edge features,it can effectively avoid the occurrence of mutational pixels on the OCT image edge results, compared to traditional edge detection algorithms.
52
Showing 1 to 3 of 3 Paper Titles