Medical Image Compression Using Packet Wavelet And Run Length Encoding
By : I Made Ari Dwi Suta Atmaja
Email : email@example.com
Faculties : Fakultas Teknik
Department : S2 Teknik Elektro
Image compression is a method that aims to reduce the use of memory, so it will facilitate the storage, processing and time of sending digital data is shorter than the data that is not compressed. This research aims to compress the medical image using wavelet packet and RLE encoding. Three types of wavelet codecs Haar, Daubechies and Biorthogonal will be used in this research. This research will compare the compression ratio, compression time and rate-distortion for each image. This research uses three threshold values of 30, 40 and 50. Experiments were conducted using five media images as data testing. The input image is test image will be processed with multi-level decomposition of PWT. Decomposition results in matrix form will be given threshold, where data smaller than threshold will be changed to zero, while the rest is not changed. The threshold result will be encoded with RLE and will be stored in a binary file * .dat format. To decompress the image, data readings begin to decode the data with RLE decoding. The result data of RLE decoding then will be reconstructed with PWT inverse filter. This reconstruction results will be displayed to the user as a decompression image. This research shown that Haar and Biorthogonal codecs give better results than Daubechies codecs in terms of image quality (PSNR) and ratio. However, for compression time, Daubechies codec is faster although not significantly. As well as for user assessment used MOS method is tested to 30 doctors, where the results of Haar codec user rating is superior to the quality of decompression images. In terms of compression ratio, Haar and Biorthogonal codecs provide better ratios than Daubechies codecs.
Keyword : Wavelet packet; Medical image; compression; Threshold; Haar codec; Biorthogonal codec; Daubechies codec; MOS.