| 摘要 |
Provided are an image processing device and an image processing method, whereby an inputted image (Din) is enlarged and a low-resolution enlarged image (D101) generated. Coefficient data (D108) for high-resolution conversion is selected according to the result of a pattern assessment of the inputted image (Din), and a low-resolution feature constituent (D102H) is converted into a high-resolution feature constituent (D103H). Whether or not the pattern of a local region of the inputted image (Din) is flat is assessed, and if the pattern is flat, coefficient data that would not result in a large change of pixel values at a high-resolution conversion unit (103) is selected. It is thus possible to carry out high-resolution conversion which is suited to being carried out in hardware, such that robustness to noise is improved while circuit scale and memory usage are reduced. |
