Image processing method, image processing device, and recording medium

An image processing device that separates an input image into a skeleton component including an edge component and a flat component and into a residual component expressed by a difference between the input image and the skeleton component, the device comprising:

• at least one non-transitory memory operable to store program code;
• and
• at least one processor operable to read said program code and configured to operate as instructed by said program code, said program code including:
• a local variation amount calculation code configured to cause the at least one processor to calculate a local variation amount that is a variation amount between a target pixel and a pixel adjacent to the target pixel;
• a skeleton component extraction filter weight calculation code configured to cause the at least one processor to calculate a filter weight for extracting the skeleton component based on the local variation amount;
• an image feature amount calculation code configured to cause the at least one processor to calculate an image feature amount of a gradient direction of a pixel value in a surrounding of the target pixel and an image feature amount of a gradient strength of the pixel value in the surrounding of the target pixel;
• a skeleton component extraction filter weight correction code configured to cause the at least one processor to correct the filter weight based on the image feature amount in the gradient direction and the image feature amount of the gradient strength;
• a skeleton component extraction filter coefficient calculation code configured to cause the at least one processor to calculate a filter coefficient based on the corrected filter weight;
• and
• a skeleton component extraction filtering execution code configured to cause the at least one processor to extract the skeleton component by applying skeleton component extraction filtering to the target pixel using the filter coefficient.