Optics: measuring and testing – By shade or color
Reexamination Certificate
2001-05-31
2004-02-03
Cherry, Euncha (Department: 2872)
Optics: measuring and testing
By shade or color
C356S405000
Reexamination Certificate
active
06687003
ABSTRACT:
FIELD OF THE INVENTION
The invention represents a novel method for recording and viewing stereoscopic images by making use of a partially separated encoding of the depth information and the colour values.
BACKGROUND OF THE INVENTION
The so-called anaglyph method for recording and viewing stereoscopic images has been known for around a hundred and fifty years. In its most common form, as applied to printed images, two mutually extinguishing filters are placed in front of the viewer's eyes and the stereogram (stereoscopic image pair) to be viewed is printed in two suitably chosen dyes, as two images brought “in register”, i.e. overlapping correctly. The filters and the dyes are complementary in the sense that, ideally, graphics printed in the one dye looks black when viewed through one filter and indistinguishable from the neutral background when viewed through the other and vice versa. Thus, the filters in front of the viewer's eyes effectively separate the two image impressions in accordance with the stereoscopic principle.
As described above, the method produces monochrome images; and it is in fact quite common to further confine the applications to line drawings and the like, where only purely black lines and curves on a neutral (“white”) background, or vice versa, are used. When either the filters or the dyes or both fail to meet the ideal requirements, a number of problems can arise, the most typical being that of cross-talk or ghosting (a colloquial term implying “ghost imaging”), where one eye receives, in addition to its “own” image, a faint copy of the image meant for the other eye. In severe cases, ghosting can interfere with stereopsis to the point of complete obstruction, especially in prolonged stereo viewing sessions. (The word diplopia, meaning “double vision”, is used for this situation; but it must be understood to mean “failure to obtain stereoscopic fusion”, rather than “observation of a double image”, since the latter can happen in stereoscopy with the perception of parallactic depth still active).
Since its discovery, the anaglyph method has been applied to line graphics, photography, motion pictures, television and computer graphics. Each of these media presents its own version of the problem of non-ideal filters and image colorants.
Quite apart from the efforts towards reducing or eliminating the ghosting in monochrome anaglyphs, several attempts have been made to extend the original method so as to allow it to give the viewer the impression of seeing a stereoscopic colour image.
Like the original anaglyph method these extensions make use of two coloured filters placed in front of the viewer's eyes. Each of the two images forming the stereoscopic pair is again displayed in colours so chosen that, ideally, each eye receives only its corresponding image, thus allowing stereopsis. However, in order that an impression of a coloured stereo image can arise, the images reaching the eyes after passing the filters can no longer appear as black on a neutral background; indeed, taken together (in the appropriate sense), they must contain all or practically all colours present in the original scene.
In the prior art, the filters chosen are approximations to ideal complementary side band filters or to an ideal mid band filter and its complementary, where the “bands” in question are ranges of the visual spectrum. Also, the two images recorded to form the stereogram are obtained essentially by filtering the original stereo pair according to this ideal of complementarity. If this filtering is performed optically, the partial images are typically passed through the same coloured filters as used for the viewing, either during the recording of the scene or as a post-processing of the images before display; if done electronically, the customary approach is to use the typical RGB-representation of the image, letting one image retain only one of these three primaries, the other only the two remaining. Either way, the two colour-filtrated partial images are then fused into a single stereogram. Thus, all colour anaglyph techniques rely on the socalled tri-stimulus hypothesis, central to colour science, that all perceivable colours within a given colour gamut can be synthesized as a weighted sum of three basic colours, negative weights requiring a special interpretation. The tri-stimulus hypothesis is assumed throughout this document, and a few remarks about halftoning, undercolour removal and black generation will be made below.
The fusion of the received colours and the correlation of the colours with the stereo image takes place in the viewer's eye-brain system; and in practice, one of the viewer's eyes will receive a much larger range of hues than the other. In the prior art, it is thus tacitly assumed that the human eye-brain system can synthesize essentially any colour, provided only that one eye receives the partial signal corresponding to the original colour's content of one primary, while the other eye receives the partial signal corresponding to the colour's content of the two other primaries.
Prior art techniques may be seen in U.S. Pat. No. 4,134,644 and U.S. Pat. No. 4,247,177.
In addition to ghosting, colour anaglyphs as produced with the prior art have been encumbered by one or more of the following problems, depending on the combinations of filters and display colours used:
The range or “gamut” of colours actually perceived by the viewer has been significantly reduced relative to that of the original scene recorded in the stereogram.
Some colours actually perceived by the viewer have shown significant chromatic deviations from the original colours in the scene recorded.
The total amount of light passing through the filters has often been significantly smaller than the light that would have reached the viewer's eyes, had no filters been applied; and, more importantly, individual colours or colour ranges have shown different degrees of lightness reduction, leading to lightness imbalances in the perceived colours.
The total amounts of light reaching each of the viewer's eyes have differed significantly, and the difference has not been correlated to luminance differences in the original scene. More precisely: owing to the colour separation, a scene colour may require of the human observer's eye-brain system an averaging of a component of a relatively high lightness seen in one eye as against a component of a relatively low lightness seen in the other. This in itself is a considerable psycho-physical challenge. If two such scene colours of approximately the same original lightness but different hues are adjacent in the image, yet appear as light and dark in opposite eyes, the two resulting averaging processes are in direct opposition.
Some colours perceived by the viewer have exhibited an undue amount of stereoscopic sheen, a characteristic lustre observable in all kinds of stereograms in homologous areas of different colours. (The strength of sheen depends in a complicated way on the original scene colour, on the colour differences between the homologous areas, on the luminance differences of the homologous areas and on the viewer's perception).
The basis of the present invention can be demonstrated by two simple experiments:
1) If a stereogram is ideally colour-separated (by electronic means, as described above) and the resulting partial images are displayed side-by-side for optical fusion, using e.g. a Brewster-type stereoscope—instead of overlapping, for fusion by means of coloured viewing filters—it is easily observed that even in this ideal case, the above-mentioned problems occur, the absolute loss of lightness owing to the absorption in viewing filters excepted, of course. In other words: the human eye-brain system is not fully capable of performing the colour fusion tacitly assumed by the prior art.
2) If, instead, the stereogram is so “colour-separated” that one partial image retains all colours while the other is converted to a gray-scale image, the latter further subjected to some reduction of contrast
Hansen Per Skafte
Sorensen Nils Lykke
Sorensen Svend Erik Borre
Birch & Stewart Kolasch & Birch, LLP
Cherry Euncha
LandOfFree
Method for recording and viewing stereoscopic images in... 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 for recording and viewing stereoscopic images in..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for recording and viewing stereoscopic images in... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3298214