Image analysis – Pattern recognition – Feature extraction
Reexamination Certificate
1999-03-02
2003-03-04
Lee, Thomas D. (Department: 2724)
Image analysis
Pattern recognition
Feature extraction
C382S118000, C382S195000, C382S203000
Reexamination Certificate
active
06529630
ABSTRACT:
BACKGROUND OF THE INVENTION
This invention relates to the technical field of a method and a device for extracting principal subjects from film recorded images and other original images which are suitable for use with photoprinters and otherwise.
Heretofore, the images recorded on photographic films such as negatives and reversals (which are hereunder referred to simply as “films”) have been commonly printed on light-sensitive materials (photographic paper) by means of direct (analog) exposure in which the film image is projected onto the light-sensitive material to achieve its areal exposure.
A new technology has recently been introduced and this is a printer that relies upon digital exposure. Briefly, the image recorded on a film is read photoelectrically, converted to digital signals and subjected to various image processing operations to produce image data for recording purposes; recording light that has been modulated in accordance with the image data is used to scan and expose a light-sensitive material to record a latent image, which is subsequently developed to produce a (finished) print. The printer operating on this principle has been commercialized as a digital photoprinter.
In the digital photoprinter, images can be processed as digital image data to determine the exposing conditions for printing; therefore, the digital photoprinter is capable of performing effective image processing operations such as the correction of washed-out highlights or flat shadows due to the taking of pictures with backlight or an electronic flash, sharpening and the correction of color or density failures and this enables the production of high-quality prints that reproduce images of the quality that has been impossible to achieve by the conventional direct exposure technique. In addition, not only the assembling of images and the splitting of a single image into plural images but also the compositing of characters can be performed by processing the image data and, as a result, prints can be outputted after various editing and/or processing operations have been performed in accordance with specific uses. Outputting images as prints (photographs) is not the sole capability of the digital photoprinter; the image data can be supplied into a computer or the like and stored in recording media such as a floppy disk; hence, the image data can be put to various non-photographic uses.
Having these features, the digital photoprinter is basically composed of an image input unit and an image output unit. The image input unit has a scanner (image reading apparatus) that reads the image on a film photoelectrically and an image processing apparatus that processes the captured image to produce output image data. The image output unit has a printer (image recording apparatus that records a latent image on a light-sensitive material by scan exposing it in accordance with the image data supplied from the image input unit and a processor (developing apparatus) that performs development and other necessary processing on the exposed light-sensitive material to produce a print.
In the scanner, reading light issuing from a light source is allowed to be incident on a film, from which projected light bearing the image recorded on the film is produced and focused by an imaging lens to form a sharp image on an image sensor such as a CCD sensor; the image is then captured by photoelectric conversion and sent to the image processing apparatus as data for the image on the film (as image data signals) after being optionally subjected to various image processing steps.
In the image processing apparatus, image processing conditions are set on the basis of the image data captured with the scanner and image processing as determined by the thus set conditions is performed on the captured image data and the resulting output image data for image recording (i.e., exposing conditions) are sent to the printer.
In the printer, if it is of a type that relies upon exposure by scanning with an optical beam, the latter is modulated in accordance with the image data sent from the image processing apparatus and deflected in a main scanning direction as the light-sensitive material is transported in an auxiliary scanning direction perpendicular to the main scanning direction, whereby a latent image is formed as the result of exposure (printing) of the light-sensitive material with the image bearing optical beam. In the processor, development and other processing as determined by the light-sensitive material are performed to produce a print (photograph) reproducing the image that was recorded on the film.
If the original image contains a human individual, his or her face is the most important element of processing with photoprinters whether they are digital or of the conventional analog type (the two types are hereinafter collectively referred to as “photoprinters”) and the quality and evaluation of the reproduced image, or print, are greatly affected by how beautifully the face of the human individual is finished.
To deal with this problem, the conventional photoprinters extract the face (face region) of a human individual from the image on a film and determines exposing conditions that enable the face to be finished in the desired way. In the case of a digital photoprinter, such exposing conditions are image processing conditions for producing output image data and in the case of an analog photoprinter, they are exemplified by the amount of exposing light and the amounts by which color filters are to be inserted.
Particularly in the case of a digital photoprinter which is capable of very flexible image processing through the processing of image data, the information on the face is important for setting image-dependent optimal image processing conditions and performing the appropriate face extraction is the way that enables high-quality prints to be produced in a more consistent manner.
Under these circumstances, various methods of subject extraction have so far been proposed for extracting various kinds of principal image subjects. To mention a few examples, skin color extraction using the color hues and saturations of an image and contour extraction by detecting the edge portions of an image are known.
However, the state of the images on films varies with the conditions under which the images were taken, such as whether they were shot with an electronic flash or backlight, or whether the films were overexposed or underexposed or whether color failure occurred. Hence, no single method of extraction that has been proposed to date is capable of ensuring consistent and precise face extraction from various states of images. The precision of face extraction can be improved by combining a plurality of methods of extracting image subjects. However, in order to perform consistent and precise face extraction from various states of images, many methods of extraction have to be combined and the time of performing the necessary processing and calculations to extract the desired principal image subject will inevitably increase to cause undesired effects such as the decrease in the efficiency of print production.
As already mentioned, the state of the images on films varies with the conditions under which the images were taken, such as whether they were shot with an electronic flash or backlight. In addition, the state of the principal subject in a particular image such as its size and orientation also varies with the image, particularly the scene in it.
In ordinary photoprinters, the subject of an image is extracted by many methods searching in four directions through the image and the results of extraction by the respective methods are combined to extract (and evaluate) the principal subject of the image, thereby ensuring that the principal image subject can be extracted with satisfactory precision from many and diverse images.
However, this requires a prolonged time to extract the desired principal image subject, which is one of the reasons for the lower efficiency of print production. A further problem with the use of many methods in extracting
Fuji Photo Film Co. , Ltd.
Lee Thomas D.
LandOfFree
Method and device for extracting principal image subjects does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and device for extracting principal image subjects, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and device for extracting principal image subjects will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3030497