An object that is reproduced by a stereoscopic display is perceived in front of the screen with depth distortion by human observers when the reproduced area of the object is truncated by a picture frame Experimental results suggest that (1) the occlusion condition, which contradicts a real three-dimensional situation (i.e, valid), causes this distortion, (2) observers, who misfuse a binocular image that includes the monocular region perceive the frontal-parallel plane as a slanted one, (3) this misperception is caused because the process of the information synthesis between the pictorial and binocular occlusion cues depend on the binocular disparity sign (i.e, it is caused when real world constraints were not satisfied) We show that adding a virtual picture frame (VFM : virtual frame) in front of the real screen enables human observers to perceive the correct depth.

View full abstract

We propose a 3D display system which can present real scenes with realistic motion parallax. In the sensing part, a scene is observed by using a camera matrix. An excellent stereo algorithm SEA recovers the depth information of the scene with the density and the sharpness required for high quality image generation. In the display part, 2D images from the viewing position of the observer are generated and presented with proper motion parallax. A novel algorithm to determine the view parameters suitable for reproducing the motion parallax on a fixed screen and an image generation algorithm which can cope with arbitrary viewing positions are described. Evaluation of the proposed algorithms has been done by using the prototype system.

View full abstract

An object swinging horizontally in a frontoparallel plane appears to move in an elliptical orbit in depth when it is observed with a neutral density filter over one eye. This phenomenon is called the Pulfrich effect. In this paper, the amount of the depth was measured as a function of pattern velocity and filter density by giving the patten a binocular disparity that cancelled the apparent depth. The magnitude of time delay given by each filter was derived from the experimental results. The apparent orbit due to a simple harmonic motion was also synthesized from the results.

View full abstract

Binocular fusional limits are measured under two conditions for observing a small stereoscopic display with a short viewing distance. One is a free-observing condition in which a subject changes his gaze point between a far and a near target alternately and estimates fusional limit at each fixating point the other is a fixating condition in which a subject fixates on a center point and estimates the fusional limit for the whole screen. In the free-observing condition, fusional limit is in proportion to the viewing angle and has weak dependence on viewing distance. In some cases the fusional limit of the far point is smaller than the disparity when the lines of sight are parallel. In the fixating condition, the fusional limit is ± 40 [arc min] as a maximum, larger than Panum's fusional limit, and has weak dependence on viewing angle and viewing distance. A method to design stereoscopic image parameters to display all objects within a subject's fusional limit is examined using the measured fusional limits.

View full abstract

An novel binocular display with accommodative compensation is experimentally manufactured for use in fatigueless 3-D displays. The observed distance is measured two ways. One combines the gaze of one eye with a depth map of the image. The other way measures the convergence point using the gazes of both eyes. The former detects the depth correctly, but is limited to the case in which a real-time depth map of the field of view is obtained. The latter produces much depth detection error, but can be applied to an arbitrary image source.

View full abstract

The ratio of the apparent size of two stereoscopic targets and the vergence of subjects viewing these targets were measured. When the two targets have small disparities within Panum's fusional zone, the ratio of apparent size conformed to the ratio of the geometrically calculated size. On the other hand, when the two targets have large disparities, subjects could not fuse two targets simultaneously, and the ratio of apparent size did not conform to the ratio of the geometrically calculated size. This can be explained by the changes of retinal disparity that depend on the vergence displacement.

View full abstract

In this paper, we propose a new framework of representing 3-D spacial information by using 'Rays'. We show that this data format includes all the information of the scene viewed from any direction, so that, one view image from an arbitrary viewpoint can be obtained by sampling ray data. We also introduce ray parameter space and examine its characteristics. We give a new definition of visual field and viewing zone and evaluate the amount of ray data in terms of ray parameter space. In addition, we show possible applications of this concept by presenting simulations, including generating an any-viewpoint image and a variable-focused image, synthesizing a real image and a computer-generated image, and compressing a hologram pattern.

View full abstract

We propose a parallel algorithm that efficiently generates an octree from a polyhedral shape representation. The algorithm is simple and applicable to general (non-convex) objects. The parallelized algorithm is based on two ideas for the assignment of multiple processors. The algorithm was implemented and tested on MIMD multi-processors having a shared memory. Experimental results demonstrate the effectiveness and efficiency of the proposed method : for example, only 2.5 seconds are required to generate an octree shape representation from a polyhedral object having 148 polygons. We discuss the performance of the proposed parallel algorithm and the features of the parallelism.

View full abstract

We discuss the uniqueness of 3-D shape recovery of a polyhedron from a single shading image, and propose an approach to uniquely determine the concave shape solution by using interreflections as a constraint. In this paper, we show that if interreflection distribution is not considered, multiple convex (or concave) shape solutions usually exist for a pyramid with three or more visible facets. However, if interreflection distribution is used as a constraint to limit the number of shape solutions (for a concave polyhedron), polyhedral shape can be uniquely determined. Interreflections, which considered to be deleterious in conventional approaches, are an important constraint for shape-from-shading.

View full abstract

In this paper, we propose a new approach to represent a set of multiview images with surface models corresponding to objects in the scene. Because the surface models overlap each other, the proposed method can cope with complex scenes in which occlusion occurs. First, a set of multiview images is partitioned into 3-D regions corresponding to the scene objects. Second, the structure of each 3-D object cor-responding to a 3-D region is recovered and represented by a surface model using triangular patches. These surface models are then placed in appropriate locations to describe occlusions. This method can be seen as an extended version of layered representation. Each layer is an uneven surface, while previous works have concentrated on planar layers. The experimental results show how the 3-D layered representation method could be applied to the next-generation 3-D image communication system.

View full abstract

Reconstruction of virtual space has become very important in various applications. Computer graphics methods based on geometric models have been studied for this purpose. To obtain more natural representation, this paper proposes a new method which represents and handles 3 D objects using the concept of “ray space.” Ray space is a space that stores the rays reflected by 3 D objects. Handling objects in real space is equivalent to getting a sliced image with an appropriate plane in the ray space.

View full abstract

This paper describes realistic, real-time 3-D reproduction of human body movements based on detection data from a small number of sensors attached to a participant for virtual space teleconferencing. A 3-D human model is created by a skeletal link model, WFM (Wire Frame Model), body segments for FFD (Free Form Deformations), and IFP (Implicit Function Primitives) for body parts In this model WFM provides fast rendering ability and FFD and IFP give natural deformation representation. A reconstruction algorithm that can deal with human movements in teleconference situations is proposed, sensed data are used to reconstruct whole body movements by utilizing motion constraints Experiments, in which four magnetic sensors for detecting the six pose parameters are attached to the upper half of a human body, show the effectiveness of the proposed 3-D human model and reconstruction method, in which human movements are reproduced at a speed of 8 frames/sec

View full abstract

One of the most useful ways to display 3-D images is with a real-time stereogram system using a holographic optical element (HOE) This simple display system presents stereoscopic 3-D images that can be without special glasses, but the quality of the images is not as good as that of images reconstructed from an optical hologram This paper shows that this problem can be solved by synthesis of a screen HOE and an optical hologram In addition, a newly developed method for making a screen HOE that is easy to synthesize is described Experimental results in this paper demonstrate the feasibility of this expressive holographic display

View full abstract

We worked on a method for extracting three-dimensional information from holographic interference fringe patterns by using an interferometer. So far, we identified the matching method by using the sum of products and the method that used an optical correlation system. In the handling that used the optical correlation system, we made a lot of masks for the depth direction and had to exchage them. In the work described in this paper, we used only one mask as complex amplitude distribution instead of the amplitude distribution, Here we present the method that extracts three-dimensional information by sifting the plane of observation phase information data with the improved optical correlation system.

View full abstract

With the shift of horizontal driving frequencies to higher frequencies, the deflection yoke inductance becomes lower, causing various problems. To avoid these problems, the voltages on both sides of the deflection yoke were raised by impressing a negative pulse on a conventional flyback pulse (positive pulse) using LC resonance circuits and switching FETs. Thus, a deflection/high-voltage integrated circuit was developed, which can drive the deflection yoke having about three times higher inductance than the conventional one. This circuit configuration made it possible to reduce the high frequency loss, leading to an about 13% reduction in power consumption.

View full abstract

An efficient edge sharpening technique for digital images is proposed. The resulting image has natural smoothness in its gray levels. The technique includes a Laplacian operator (a traditional technique for image enhancement) and the phase-only synthesis in Fourier analysis. The adaptive weighted functions are the S-function and π-function, often used in fuzzy theory. Furthermore, a human visual property is also introduced to attain effective results for human eyes. The processed image results were evaluated using the cross section of the gray levels of the image. Simulation examples indicate that the proposed technique is more effective than the traditional one.

View full abstract

We evaluateded a human self-control system for wide-visual field stereoscopic images using measurements of body sway and eye movement caused by vection. First, it was found that the locus length of body sway did not depend on the degree of the moving rectangular target with a uniform disparity. We used the same target with the same disparity in one period and zero disparity in the another period. The body sway was measured while the part of the target with disparity was moved repeatedly side to side and the other part was stationary. The locus length of body sway had a longer value for the moving target behind the stationary one than in other cases. These results agree with the results of the subjective assessment in the psychological experiment which used real targets. Moreover, the frequency of OKN (optokinetic nystagmus) was almost the same for the uniform disparity target in the measurement of eye movement. It was also found that OKN occurred frequently with a moving target behind a stationary one. However, OKN was almost nonexistent in other cases. In conclusion, we examined the possibility of the objective evaluation of a human self-control system based on these results.

View full abstract

Hi-Vision products have become very popular as the technology has been developed. Consequently, there is a growing demand for a Hi-Vision video camera View-Finder. We have developed a monochrome CRT that can be used as a Hi-Vision View-Finder. In this paper, we will discuss the key technologies of this CRT.
(1) Steps taken to improve resolution.
(a) Improved glass bulb which can take a high voltage.
(b) Improved phosphor screen.
(c) Improved electron gun.
(2) Steps taken to improve circuitry. We have developed :
(a) A wide band video circuit, a stable deflection circuit, and a high-voltage circuit.
(b) A precision beam correction method based on a unique digital circuit.
We have achieved our target specification for this prototype and have confirmed that the CRT can be used for Hi-Vision View-Finder applications.

View full abstract