Hybrid image compression using VQ on error image
Prachi Natu, Shachi J. Natu, Tanuja Sarode · 2017
Here color image is compressed using vector quantization (VQ) on transformed coefficients. First, image is compressed using hybrid wavelet transform to achieve high compression ratio. Error image is obtained after transform based compression. Then VQ is applied on compressed image to increase compression ratio. Codebook of size 32 is generated. As codebook size is small, it gives higher compression ratio. Distortion at such high compression ratio is calculated. Further, to reduce distortion, VQ is applied on error image too. This compressed error image is added to VQ compressed image and reconstructed image is generated. It reduces distortion by nearly 10% as compared to previous case. Proposed algorithm is tested on images of different classes. It gives much better compression at the cost of computational complexity.