Image Processing Of Sub-sampling And Super-sampling

• Main Authors: Chen, Chiahong, 陳嘉宏
• Other Authors: Jeng, Jyhhorng
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/14783243866675632976

碩士 === 義守大學 === 資訊工程學系 === 100 === Today, most images and videos are stored in various formats and image resolutions. Images are displayed on several multimedia terminal devices at different levels of image resolution such as personal computers, personal digital assistants (PDA), smart phone, and digital TV. The image size must be scaled to fit the device resolution or the viewer’s preference. Image resizing is also necessary for video display at different image resolution in order to enhance the image quality. Image resizing technique to reduce the spatial resolution of image or video content before delivery to these mobile devices is needed. In this work, we apply discrete cosine transform (DCT) method to transform the spatial domain to frequency domain. Many resizing algorithms are based on the DCT domain. Interpolation of an image is known to be more accurate and efficient in the DCT domain than in the spatial domain. We used sub-sampling and super-sampling to resize the image resolution. Hence, we analyze coefficient values to distribute in the frequency domain. Furthermore, we propose an arbitrary scaled method to resize image in the DCT domain. Finally, we propose multi-stage of super-sampling and mix domain methods for super sampling. Experimental results demonstrate that the proposed method can achieve in the super-sampling, and have better result than the relevant existing methods.