Adaptive Image Interpolation Based on Local Edge Directions

• Main Authors: Hung-Chu Wu, 吳鴻居
• Other Authors: Yen-Tai Lai
• Format: Others
• Language: en_US
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/83736231702749492449

碩士 === 國立成功大學 === 電機工程學系碩博士班 === 97 === In recent years, image/video format conversion becomes increasingly important due to the rapid development of digital display devices. Image/video format conversion includes two main fields: image scaling and de-nterlacing, which both can be performed by image interpolation. Conventional image interpolation methods suffer blurring problems in edge regions, so it is hard to produce sharp and clear visual effects. The objective of our research is to find an interpolation method which can provide good image quality while keeping the edge sharp. In this thesis, we research on two applications of image interpolation, image scaling and de-interlacing. We propose the corresponding interpolation algorithms based on edge directions to classify the image/video into edge regions and non-edge regions for adaptive interpolation. For de-interlacing, motion detection process is adopted to divide the video into moving regions and static regions. Good video quality can be achieved by choosing the proper interpolation method according to the different local features in video sequence.