Capsule endoscopy image coding using wavelet-based adaptive vector quantization without codebook training

The amount of image data generated by capsule endoscopy is so large that data compression is desirable. In our compression scheme, a codebook (CB) replenishment mechanism is incorporated in a wavelet-based adaptive vector quantizer (VQ). The progressive SPIHT coding is used to meet the quality demand from a user. Furthermore, in our VQ implementation, a modeling, rather than a training, technique is proposed to generate an initial CB, where a pseudo-noise sequence is used to create such a CB at both the encoder and the decoder. Experimental results show that the proposed CB modeling method produces comparable performance to the one using CB training.