Neural Net Classification and Low Distortion Reconstruction to Halftone Images

• Main Authors: Yu, Che Sheng, 游哲生
• Other Authors: Pao-Chi Chang
• Format: Others
• Language: zh-TW
• Published: 1997
• Online Access: http://ndltd.ncl.edu.tw/handle/84929851886191587108

碩士 === 國立中央大學 === 電機工程學系 === 85 === The objective of this thesis is to reconstruct gray-level images fromhalftone images. We develop high performance halftone reconstruction methodsfor several commonly used halftone technigues. For better reconstructionquality, image classification based on halftone techniques is placed beforethe reconstruction process so that the halftone reconstruction process canbe fine tuned for each halftone technique. The classification is based onsimplified one-dimensional correlation of halftone iamges and processedwith neural networks. The classification method reached 100% accuracy inour experiments. For image reconstruction, we develop LMS adaptive filter(LMS) method, minimum mean square error (MMSE) method, and hybrid method.The hybrid method yields best reconstruction image quality and highprocessing speed. In addition, the LMS method generates optimal imagemasks which can then be applied to MMSE and hybrid methods to setup optimalreconstruction tables.