Wavelet-based Image Analysis

• Main Authors: Jun-Wei Hsieh, 謝君偉
• Other Authors: Kuo-Chin Fan and Hong-Yuan Liao
• Format: Others
• Language: en_US
• Published: 1995
• Online Access: http://ndltd.ncl.edu.tw/handle/31376487557834950269

博士 === 國立中央大學 === 資訊及電子工程研究所 === 84 === This dissertation discusses several related research topics about wavelet-based image analysis. We try to apply wavelet theory to several different, but related image analysis problems such as edge detection, image registration, and shape from shading. This research is extremely valuable in the fields of computer vision, image processing, pattern recognition, automatic navigation, robot vision, and other related areas. Except for a general introduction and conclusion, the dissertation mainly discusses the following three topics about wavelet-based image analysis problems. They are: (1) a new wavelet-based edge detector, (2) a new wavelet- based image registration technique, and (3) a new wavelet-based shape from shading technique. The relations among the three works are reported below. Basically, the proposed image registration technique is an edge-based approach. Therefore, the work of the wavelet-based edge detection can support its need. On the other hand, when the image registration task is done, the proposed shape from shading technique can recover the depth of every registered feature point and thus facilitate the subsequent high level image analysis tasks. Experimental results are reported for each topic to corroborate the proposed theories.