| Summary: | 碩士 === 國立成功大學 === 電機工程學系 === 87 === An Adaptive 3-D Discrete Cosine Transform Coder for Medical Image Compression
Chang-Wei Lin*, Shen-Chuan Tai**
Department of Electrical Engineering
National Cheng Kung University, Tainan, Taiwan, R.O.C.
Abstract
Due to the bandwidth and storage limitations, medical images must be compressed before transmission and storage. However, the compression will reduce the image fidelity, especially when the images are compressed at lower bit rates and block-coded techniques are used, the reconstructed image will generate blocking artifacts and the image quality will be severely degraded.
2-D compression techniques achieve better energy compactness than 1-D types in terms of transform coding. Our method is not only based on that assumption but is also extended to 3-D for better compression efficiency in terms of energy compactness. A segmentation technique based on the local energy magnitude was developed as a criterion to segment sub-blocks of the image into different energy levels. Then, those sub-blocks were gathered with the same energy level to a 3-D cuboid. Finally, 3-D Discrete Cosine Transform (DCT) was employed to compress the 3-D cuboid individually. Conventional transform coding techniques compress images in a block by block manner, therefore, blocking artifact occurred. Our strategy, however, takes the inter-block correlation into account by using the segmentation results. This could reduce the blocking effect and raise the compression ratio by its excellent energy compactness in 3-D transformation.
We tested our 3-D DCT image coder on different medical images. Test images include X-ray, sonogram and angiogram images. The sizes of them are all 512’512 with 8 bits per pixel monochrome images. The reconstructed images of those test data achieve the bit rate lower than 0.25 bits per pixel (bpp) whose compression ratios are higher than 35. In addition, we used the most widely used compression tool of JPEG and other published technique to compress the same test images. The qualities of decoded images by JPEG and the other strategy demonstrated the high performance of our proposed method in terms of compression ratio and image quality.
*The Author **The Advisor
|
|---|
