Image analysis – Applications – 3-d or stereo imaging analysis
Reexamination Certificate
1995-08-04
2001-06-26
Patel, Jay (Department: 2623)
Image analysis
Applications
3-d or stereo imaging analysis
C345S422000, C382S285000
Reexamination Certificate
active
06252982
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to a method and to an image processing system for processing a 2-dimensional input image to create a 2-dimensional output image with enhanced depth sensation.
2. Background Art
A particular problem pursued in the field of image processing is electronic image enhancement to give the viewer the impression of a 3-dimensional (3D) scene when watching an image on a 2-dimensional (2D) display. With the advent of virtual reality, the search for a solution to represent or suggest depth in electronically generated visual 2D information has gained momentum.
A known way to provide depth sensation is based on stereopsis, i.e., stereoscopic vision. Stereopsis is evoked by two images of the same scene as perceived from two positions a little way apart. One of the images is observed by the left-hand eye, the other one is observed simultaneously by the right-hand eye. Discrimination between the left-hand and right-hand images is brought about by auxiliary viewing devices, for example, spectacles with left-hand transmission properties different from the right-hand ones.
Alternatively, the brain can reconstruct 3D-images when the right and left fields of vision are presented in rapid succession. To this end, a series of alternatingly clear and opaque strips are made to oscillate so fast that an image on a screen behind the oscillating strips appears clear. A special camera records the two alternating images. As the slits and the strips lock together and scan quickly across the screen both images are assembled by the brain of the observer to form a single 3D image. Under ideal circumstances, the 3D effect is independent of the distance between the observer and the screen. See, e.g., New Scientist, Feb. 1994, page 19; Time, Oct. 18, 1993, pp. 64-65, or Business Review Weekly, Feb. 7, 1994, pp. 44 and 48. This approach requires that the right-hand and left-hand information fields of vision be separately available and also needs a dedicated presentation system.
Another manner to perceive depth in 2D images is based on so-called Single Image Random Dots Stereopsis, also referred to as autostereopsis, i.e., adaption of the convergence of the eyes to a pair of two superimposed stereoscopic images. Two superposed stereoscopic images give the sensation of depth by skew-eyed watching, as convergence and accommodation of the eyes are decoupled. This is difficult to learn, and once learned it is experienced as fatiguing on the long run.
OBJECT OF THE INVENTION
It is an object of the invention to provide an alternative system for the creation of images that give 3D sensation based on the processing of 2D image data. It is another object to provide a system that is able to operate on conventional 2D images, such as TV pictures.
SUMMARY OF THE INVENTION
To this end, the invention provides a system as specified in the preamble, characterized by an identificator operative to select, according to a predetermined criterion, at least one area being a portion of the input image; and a processor operative to create the output image by changing a property of the area relative to a corresponding property of a complement of the area in the input image. The depth enhancement can be achieved through, e.g., selectively processing the area, or its complement or complementarily processing both the area and its complement. The processor is operative to modify, e.g., contrast, hue, brightness or sharpness.
The invention is based on the insight that in essence human vision is spatially non-uniform, and that depth sensation can be made more pronounced by treating the input image non-uniformly. For instance, a person looking straight at an object perceives that object as being clearer, sharper, or more pronounced than other objects lying at the same distance from that person but whose images occur in the periphery of the person's field of vision. The system underlying human vision differs from artificial image-processing systems. A camera, for example, does not discriminate between objects that lie substantially on the camera's optical axis and objects that do not. Actually, the camera's designer will do everything to eliminate such non-uniformities, that derive from spherical aberration, from the mapping of the image onto the film. The invention therefore advocates an approach for artificial image-processing essentially different from that used heretofore.
For example, by enhancing the contrast in the area of the main object shown by the input image relative to its surroundings, depth is felt to be more strongly present. Similarly, brightness and/or hue can thus be used to augment depth sensation. Also, the area's complement can be treated by reducing contrast, brightness or hue to stress the content of the area itself. Alternatively, selective filtering out of higher-frequency components from the area's complement increases the area's sharpness and contributes to the area being presented as more pronounced. Other quantities may be used, or combinations of aforesaid quantities, as depth cues in order to non-uniformly process the input image.
Alternatively, or supplementarily, the information content of the scene represented by the input image is identified on the basis of, e.g., the contours of an object, distribution of colours, luminance and polarization, coherence in case of movement, etc. The information contents may be used to select the area for processing as specified above. In a further alternative embodiment, the area may be identified on the basis of object recognition.
The invention is inspired by developments in the art of painting, wherein new techniques (e.g., the use of perspective, pointillism) and novel manners of expression (e.g., pop-art, or non-figurative art) each time have taught fresh ways of perception, that have been accepted by the audience only gradually. According to the invention, particular characteristics of the 2D image, called the monocular depth cues, are modified so as to control depth sensation. In order to appreciate the accomplishment of depth generation, the spectator may have to learn anew how to look at the novel depth-enhanced 2D images generated by the system of the invention.
Preferably, the system in the invention is suitable for handling moving pictures on a real-time basis. Preferably, the processor is user-controllable to individualize or customize the representation of the output image. For example, a conventional TV set can be provided with a system in accordance with the invention. A user may then adapt a TV broadcasting to his personal liking.
Note that the invention provides advantages not only in the consumer field, such as video entertainment, but also in the field of professional image analysis and processing, such as in medical applications (NMR, X-ray scan), in photographic art, and computer aided design. The invention also may find a use in multi-media systems that are designed to combine in a compound image, under user-control, information items obtainable from a large variety of information sources, so that the depth sensation of the compound image can be freely manipulated.
The invention also relates to a method of processing a 2-dimensional input image to create a 2-dimensional output image with enhanced depth sensation, characterized by selecting, according to a predetermined criterion, at least one area being a portion of the input image, and changing a property of the area relative to a corresponding property of a complement of the area in the input image. At least contrast, hue, brightness or sharpness the input image can thus be changed non-uniformly. Digital signal processors of a general-purpose type can be used to implement the method of the invention.
REFERENCES:
patent: 4661986 (1987-04-01), Adelson
patent: 4991224 (1991-02-01), Takahashi et al.
patent: 5113213 (1992-05-01), Sandor et al.
patent: 5247590 (1993-09-01), Fukuhara et al.
patent: 5295199 (1994-03-01), Shino
patent: 5309522 (1994-05-01), Dye
patent: 5363475 (1994-11-01), Baker et al.
pate
Collet Marnix G.
Haisma Jan
Widdershoven Franciscus P.
Patel Jay
U.S. Philips Corporation
Verdonk Peter
LandOfFree
Image processing system for handling depth information does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image processing system for handling depth information, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image processing system for handling depth information will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2452171