Home About Login Current Archives Announcements Editorial Board
Submit Now For Authors Call for Submissions Statistics Contact
Home > Archives > Volume 20, No 8 (2022) > Article

DOI: 10.14704/nq.2022.20.8.NQ44909

A Hybrid Discrete Wavelet Transform with Vector Quantization for Efficient Medical Image Compression

Mohammed Fadhil Radad , Ali Obeid and Ali Al-Fayadh


The prevalence of chronic diseases, sadly, has skyrocketed in recent decades. As a result, there has been a sharp rise in the frequency with which medical imaging is used for diagnosis. To help doctors make quick, reliable diagnoses, numerous imaging technologies and software packages have arisen. Therefore, researchers in this sector are faced with the formidable task of medical image compression in order to reduce the storage capacity required for these images. In addition, reducing the size of an image through compression makes it much simpler to transfer the image over a network. To improve medical image compression, a hybrid approach between the Wavelet Transform Technique (DWT) and Vector Quantization (VQ) is proposed here.The method of compressing the medical image that has been proposed seeks to achieve a high compression ratio while preserving the diagnostic information of the image. To begin, the noise in medical images caused by splash, salt, or any tiny particles is reduced while an edge is preserved. This is done while keeping the edge. After applying DWT to the images, which is a lossless compression method for the wavelet coefficients in the lowest frequency sub-band, the images were further compressed. However, the thresholding approach was used to generate coefficients for the high-frequency sub-bands because it was the most effective. As a consequence of this, the result was given a vector quantization by the utilization of the back propagation Neural Network (BPNN) technique. The architecture of BPNN is known as a Feed-Forward Neural Network. Any task requiring a close approximation and this architecture type will do the trick. Specifically, this Neural Network employs a rule for learning from error. The use of an Artificial Neural Network architecture makes for a highly effective technique for image compression (ANN). Compression ratio (CR), Peak signal-tonoise ratio (PSNR), and mean square error (MSE) are all examples of different types of metrics (MSE). Any compressed image can be evaluated based on a set of standards.The proposed method is able to enhance compression performance and achieve a good compromise between compression ratio and image visual quality.


Hybrid Medical Image Compression, Discrete Wavelet Transformation, Back Propagation, Compression ratio, Peak Signal to Noise Ratio, Mean Square Error

Full Text