Progressive Image Transmission based on Quadtree Segmentation

• Main Authors: Yi-Hsuan Chiu, 邱宜萱
• Other Authors: Yu-Chin Hu
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/99916312203415279108

碩士 === 靜宜大學 === 資訊管理學系研究所 === 92 === In this thesis, we propose two Progressive Image Transmission methods based on quadtree segmentation technology. To understand the affect of highly pixel similarity in the digital images, we use three layered quadtree structure in sender of the transmission. We use quadtree segmentation technique for image segmentation because its flexibility. It divides input image into variable sized image blocks, and then different compression strategies to deal with these image blocks of different sizes. By selecting suitable thresholds in the quadtree segmentation technique, several image hierarchies can be thus generated. In our first method, we integrate block truncation coding to improve the image quality of compressed image. In our second method, we combine vector quantization for reduce the required bit rates. Under the low transmission bit rate, use these two methods can let receiver gain rough image in the first few rounds and better reconstructed image at the final round. Based on the results of simulations, the total transmission bit rate is lower than bit plane method and linear prediction method. The number of rounds to be transmitted can be adaptively selected according to the user requirement.