Skip to main content

Adaptive Wavelet Packet (AWP) Image Coding

Wavelet packets are an effective representation tool for adaptive waveform analysis of a given signal. Our research has investigated the marriage of wavelet packet representation with zerotree quantization for image coding purpose. In our ICIP'1999 paper, we introduced the notion of a general zerotree structure, termed as the compatible zerotree, which can adapt itself to any arbitrary wavelet packet basis. An efficient coding algorithm, the compatible zerotree quantization (CZQ) algorithm, was developed based on this structure which compared favourably to the then state-of-the-art image coding algorithms. The ICIP'2001 paper proposed a new wavelet packet basis selection paradigm which emphasized the crucial role of the quantization strategy employed by the image coder. This paradigm was coupled with a novel Markov chain based estimation of the cost of zerotree quantization to develop a progressive wavelet packet image coder which gave better results than its wavelet counterpart. The TIP'2003 paper detailed the construction of compatible zerotrees and combined this with a fast and progressive adaptive wavelet zerotree image coder yielding favourable subjective quality as compared to the then state-of-the-art image coders.

CZQ comparative results 

In a separate work that appeared in a DCC'2003 paper, we investigated the combination of wavelet packet basis selection with stack-run quantization strategy for image coding. The SRWP coder works by selecting an optimal wavelet packet basis for the given image and encoding the quantization indices for significant coefficients and zero runs between coefficients using a 4-ary arithmetic coder. Due to the fact that the SRWP coder exploits the redundancies present within individual subbands, its addressing complexity is much lower than that of the wavelet zerotree coding algorithms. Experimental results show coding gains of up to 1.4dB over the benchmark wavelet coding algorithm.

 

SRWP image coding results

Relevant Publications

  • NM Rajpoot, RG Wilson, FG Meyer, RR Coifman,
    Adaptive wavelet packet basis selection for zerotree image coding,
    IEEE Transactions on Image Processing, 12(12):1460-1472, December 2003
  • AM Awan, NM Rajpoot, SA Hussain,
    Stack-Run Adaptive Wavelet Image Compression,
    in Proceedings IEEE Data Compression Conference (DCC'2003), March 2003
  • NM Rajpoot, RG Wilson, FG Meyer, RR Coifman,
    A new basis selection paradigm for wavelet packet image coding,
    in Proceedings International Conference on Image Processing (ICIP'2001), Thessaloniki (Greece), October 2001
  • NM Rajpoot, FG Meyer, RG Wilson, RR Coifman,
    On Zerotree Quantization for Embedded Wavelet Packet Image Coding,
    in Proceedings International Conference on Image Processing (ICIP'1999), Kobe (Japan), October 1999