Abstract

In this paper three filter bank structure are evaluated by Field Programming Gate Array implementations. The traditional non polyphase structures, traditional polyphase structure and lifting structures are three filter banks. The compression performance are examined by three filter structure. For the filter coefficients, optimal quantized values are found for each structure. Discrete Wavelet Transform codec generates the best possible Peak signal to noise ratio performance for a given structure using these coefficients, in Discrete Wavelet Transform we use filter bank instead of using a single filter. By evaluating the performance optimal choices can be made for a biorthogonal 9/7 Discrete Wavelet Transform implementation based on the given application. After quantization here filter bank properties are preserved and not the properties of single filter. In this traditional Discrete Wavelet Transform can be implemented in multiple ways and then it introduces figure of merit examined hardware implementation. Two quantization techniques are used to better the performance of quantized filter bank to get higher throughput filter bank in polyphase form is used than non polyphase structure. In lifting structure, six different optimal quantized lifting implementations are designed and evaluated. To accept the proposed scheme, for the 2-dimentional Discrete Wavelet Transform computation this circuit is designed, simulated, and implemented in Field Programming Gate Array.

Authors and Affiliations