SVD壓縮照片技術結合超解析度插值方法的探討

SVD Compression Technique Combined with the Super-Resolution Interpolation Method

壓縮照片技術可以幫助照片輕量化，超解析度方法則能夠提昇照片的解析度大小，本方法希望將這二者的領域結合起來，也就是說既想要壓縮照片的內容，也想要同時放大照片的解析度品質，在壓縮照片時，以不要太過破壞照片內容細節的前提下，也能保留原始照片中的良好輪廓，使照片輕量化，並提升照片的解析度，本方法在超解析度方面，將使用插值領域，並列舉出最近鄰插值、雙線性插值、雙三次插值這三種插值方法來依序對照片進行解析度提昇，並評測其效果，而壓縮照片的方法將選用 SVD 壓縮技術，來與超解析度插值方法結合，且使用 PSNR 為評價指標。

The image compression technique and super-resolution method can enhance light-weight images and scale the image resolution, respectively. In this research, these two methods were combined to compress the image content and up sample the image resolution quality. When the image is compressed, the original image contour is retained under the “restrict” condition to prevent the image content from being excessively damaged. This process allows for enhancing of the image light weight and scaling of the image resolution. In this study, nearest-neighbor interpolation, bilinear interpolation, and bicubic interpolation were applied in the super-resolution domain to scale image resolution. An SVD technique was employed to compress the images and combine the three interpolation methods. This study used the PSNR to measure the processed image quality.