Image Restoration Based on Wavelet Semi-soft Threshold Transform and BP Fuzzy Neural Network

摘要：

　　Image restoration aimed to recover the original image to from degraded images and degenerate function.Fuzzy logic systems and neural network can complement each other quite well.In this paper,a novel Image Restoration approach is developed.Wavelet Semi-soft Threshold Transform and of our method is utilized to image restoration.Firstly,Wavelet Semi-soft Threshold Transform method is used to image denoising.Then,the image is classified into several regions using fuzzy sets,which are smoothing,texture and edge regions to obtain the input of BP Fuzzy Neural Network.Sliding window is used to extract features and input the training data.Finally,the output of BP Fuzzy Neural Network is the restored image.

关键词： Image restoration Wavelet Semi-soft Threshold Transform Fuzzy BP Neural Network Sliding window

作者: Wenjing Pei Yingmin Jia

作者单位: The Seventh Research Division and the Center for Information and Control,School of Automation Science and Electrical Engineering,Beihang University(BUAA),Beijing 100191,China

会议类型: 国际会议

会议名称: 2019中国智能自动化大会(CIA,2019)

会议地点: 江苏镇江

会议语种:英文

页码: 620-628

在线出版日期: 2019-09-20（万方平台首次上网日期，不代表论文的发表时间）