Title: An image detail enhancement of smart product UI interface based on stationary wavelet transform

Authors: Geng Chen; Quzhi Huang

Addresses: Department of Information Technology, Concord University College Fujian Normal University, Fu'zhou, 350117, China ' Department of Information Technology, Concord University College Fujian Normal University, Fu'zhou, 350117, China

Abstract: To overcome the problems of low image segmentation accuracy, low image signal-to-noise ratio and long image enhancement time in traditional methods, an image detail enhancement method of smart product UI interface based on stationary wavelet transform is proposed. The Gaussian mixture model is used to obtain the image parameters of the UI interface of smart products, and the image of multiple pixels is divided into marked categories by the maximum posterior probability criterion, so as to realise the segmentation of image noise area and normal area. The two-dimensional stationary wavelet transform is performed on the noisy area, and the inverse stationary wavelet transform is performed on the stationary wavelet coefficients to obtain a reconstructed image with enhanced details. Experimental results show that the image segmentation accuracy of this method fluctuates in the range of 96%-98%, the signal-to-noise ratio is 55.3 dB, and the average image enhancement time is 66.9 ms.

Keywords: stationary wavelet transform; smart products; UI interface; image detail enhancement; Gaussian mixture model; image reconstruction.

DOI: 10.1504/IJMTM.2024.137388

International Journal of Manufacturing Technology and Management, 2024 Vol.38 No.1, pp.66 - 80

Received: 26 May 2022
Accepted: 01 Sep 2022
Published online: 15 Mar 2024 *