Contourlet transform based subband normalization for region based medical image compression

Abstract

Low bit rate compression approach has been proposed for easy transmission of medical image from one place to another. Contourlet transform technique is used for obtaining multi directional subbands and for capturing the fine details of the image. Transformed coefficients are then normalized using a mathematical approach for each subbands, followed by quantization and encoding. Arithmetic coding techniques are used for entropy encoding. Experimental results shows that the visual quality as well as the compression ratio of the reconstructed image is high compared to the existing wavelet based compression technique at low bit rates. Contextual compression is well supported for such conditions where the unwanted regions are compressed lossy, but with good visual quality. The result of the proposed method shows better compression performance with good visual quality at low bit rates.