ABSTRACT 

Digital images in their uncompressed form require extremely large amount of storage capacity that needs large transmission bandwith for their transmission over networks. Image compression is the process of removing redundant information from the image so that only essential information can be stored in order to reduce the storage size, transmission bandwidth and transmission time. This research work includes different image compression techniques used for image compression. Comparative analysis of image compression was carried out using different image compression techniques. The different image techniques are Embedded Zero Wavelet Tree (EZW), Discrete Wavelet Transform (DWT) and Set Partition in Hierarchical Tree (SPIHT). MATLAB programs were written for each of the image compression techniques. The results obtained showed that set partition in hierarchical tree with three dimensions (SPHIT_3D) technique produced the best image compression with the highest value of 43.6dB as peak signal to noise ratio (PSNR) and 2.838 as the lowest mean square error (MSE) while Embedded Zero Wavelet tree (EZW), level 5 produced lower peak signal to noise ratio (PSNR) value of 14.29dB and the highest value of MSE of 2423. Therefore, (SPHIT_3D) technique is proposed in this research work because it is more efficient than the other image compression techniques and has the least value of mean square error (MSE), highest value of peak signal to noise ratio (PSNR), better image quality value and best recovery in comparison to other image compression techniques in all given images formats. The study further recommended that since image compression is an important tool in many digital outfits, there is need for improvement in the compression of images so as to obtain an output of quality almost the same as the original image coupled with a reduced image size among other recommendations. This research work has contributed immensely in image compression by reducing the storage size of the image used in this study and producing better compressed image. This was achieved using compression ratio, mean square error and peak signal to noise ratio.