The Visibility of DCT Quantization Noise: Spatial Frequency Summation
作者： A. B. Watson , H. A. Peterson , A. J. Ahumada , Irving C. Statler
DOI:
关键词: Quantization (signal processing) 、 JPEG 、 Discrete cosine transform 、 Vector quantization 、 Image compression 、 Noise shaping 、 Data compression 、 Computer vision 、 Trellis quantization 、 Artificial intelligence 、 Algorithm 、 Computer science
摘要: Computational models of the ability to detect image compression artifacts facilitate optimization parameters, such as JPEG quantization matrix. For simplicity, some assume that visibility containing different spatial frequency components is determined by most visible component, is, no summation over components. Using type noise generated in Discrete Cosine Transform (DCT) domain, we find a degree between probability and contrast energy summation.
jaxa.jp参考文章(0)