Resolution Enhancement via Discrete Wavelet Transformation

The patent segments the correction process into two distinct stages: first applying simple filter correction to handle low-frequency components, then performing addition processing to restore high-frequency components. This segmentation allows each processing stage to focus on specific frequency ranges, achieving both computational efficiency and complete information restoration.

The patent implements continuous useful action by sequentially applying filter correction followed by high-frequency restoration without discarding intermediate results. The correction unit preserves the filtered image data and adds the high-frequency components back, ensuring that no useful information is lost in the correction process while maintaining processing efficiency.

The patent applies local quality by performing correction processing on local areas independently. The correction unit divides the image into multiple local regions and applies correction to each region based on its own characteristics, allowing the processing to adapt to local variations in correlation between images and preventing overcorrection in regions where correlation is weak.

The patent uses partial action by selectively applying correction only to regions where it is beneficial. The correction unit evaluates each local area and applies correction processing only where needed, avoiding excessive correction in regions where the simple filter correction is already sufficient, thus preventing unnatural artifacts while maintaining productivity in regions that benefit from enhancement.