Discipline: Computer Sciences and Information Management
Subcategory: Computer Science & Information Systems
Matthew Alston - Benedict College
Our project is to compress an image using DCT (Discrete Cosine Transform). The goal of any compression algorithm is to reduce the data as much as possible without undue loss of information. Other considerations include simplicity, speed of computation and flexibility. Our method is fairly simple. We use a single image to ensure that our results are not skewed by the quality of the original picture. We manipulate different aspect of the image compression algorithm such as the scaling of the quantization matrix and DCT coefficients. After compressing the image, we calculate the distortion and the compression ratio of the given image. The distortion indicates how much the image has been deteriorated by the compression technique. One key finding that has appeared in our results is that the distortion of the image is tied to the scaling parameter. Compression ratio and distortion are directly related. The scaling parameter and the distortion seem to be inversely linked so that as one rises the other falls. Our conclusion is that the modification of the parameters can lead to a least distortion and least compression ratio. Different stages of the compression are shown using MATLAB software. Our future work will be to lessen the distortion while improving the compression by lowering and raising different parameters of the algorithm to get better results.
Funder Acknowledgement(s): Dr. Samir Ray Choudhury supported us by the NSF grant #1436222
Faculty Advisor: Naima Nhaeed, naheedn@benedict.edu
Role: I collected all the data using MATLAB, according to my Professor's instruction. I made all the graphs.