二維離散小波轉換晶片之設計

Abstract

隨著網路的普及率不斷的升高，越來越多的人使用網路。與多媒體的結合，聲音與影像更使得網路的世界增添了不少的色彩。然而影像與聲音的大量資料卻造成了傳輸的速度緩慢，使用者常常要浪費不少的時間在等待上。因此如何減少資料量是一個很重要的話題。 影像的壓縮是目前很熱門的領域，目前在靜態影像的壓縮上，主流是採用JPEG的規格，然而面對越來越高品質影像的需求，JPEG的壓縮倍率率卻只能侷限在20~40之間，同時又有方塊效應的存在，因此一種更有力的壓縮方式被提出---離散小波轉換和零數編碼。但是離散小波轉換計算量相當龐大，在面對即時影像處理的應用上，即感不足，因此這裡提出了一個二維離散小波轉換的硬體架構，並設計出一顆高性能的晶片，在即時影像處理的系統上擁有很好的效果。

Nowadays , more and more people know the internet and use it . The world of internet has more funny with multimedia which has image and voice. However, it is slow that transmitting a large amount image and voice data. The users of internet may waste much time to wait. So decreasing the amount of data is a very important topic. The image compression is a hot topic nowaday. The standard of the still image compression is JPEG, but the compression ratio and quality of JPEG are not good enough. The compression ratio of JPEG is between 20 to 40.It also has other disadvantages, such as block effect. The more powerful method for image compression is proposed "Discrete Wavelet Transform and Zerotree". The calculation of DWT is very large, so it is difficult for real-time image processing. The architecture of a fast 2-D DWT hardware is proposed and it provides high performance for real-time image processing.