Improving Super Resolution Algorithm by Adaptive Segmentation and Ridge Regression

Abstract

The application of image super-resolution technologies in recent years has increased noticeably The main purpose of super-resolution is to generate high-resolution (HR) images from low-resolution (LR) images In this Thesis an efficient SR algorithm is proposed Multiple linear regression models and ridge regression models are established with sixteen oriented details from LR images by the designed filters Afterward the two reconstruction models are utilized respectively to estimate global and local details of HR images with the corresponding oriented details that are acquired from corresponding preliminary HR images by the same filters For more adaptively utilizing the straight line segments characteristics in an image Canny line detection and Hough transform are applied to build local reconstruction model Experimental results show that the proposed algorithm produces HR images with better in both the visual quality and the objective measurements

Date of Award	2016 Aug 8
Original language	English
Supervisor	Shen-Chuan Tai (Supervisor)