DIGITAL IMAGING DEVICE USING BACKGROUND TRAINING

PROBLEM TO BE SOLVED: To provide a method for a digital imaging device that can discriminate an object image from a background. SOLUTION: Various backgrounds, shapes and templates are scanned and a feature is extracted and a memory stores the feature. When a composite image (total of an object image and background) is scanned, the feature is extracted from the composite image. Correlation of the feature in the composite image is taken from that of the stored feature of the background, and which background exists is identified. If necessary, the feature of the background from the memory is shifted two-dimensionally, to adjust the displacement of the background. The image data corresponding to the background are deleted from the composite image. This invention can be applied also to extract information written on a blank paper sheet.