The multi-view stereoscopic display has several merits such as requiring no special glasses and providing motion parallax. Since existing displays have flat screens, observers can not watch images from behind. If a multi-view stereoscopic display which has a cylindrical screen is developed, it would enable observers to watch the stereoscopic image like a multiplex hologram from all the horizontal directions around it. But it is difficult to make a cylindrical multi-view stereoscopic display with conventional technology such as a lenticular sheet and two-dimensional spatial light modulators such as LCD or CRT. Therefore we propose a new type of three-dimensional display, which has a cylindrical parallax barrier and a one-dimensional light array which consists of semiconductor light sources in a vertical line. The light array rotates along the cylindrical parallax barrier and its intensity is modulated synchronously. We have developed an experimental display and we confirmed that we can watch different images from all the directions surrounding.
Holography is an ideal technique for 3 D displays. However, since precise modulation of light waves is required, it has long been very difficult to achieve electro-holography. Recently, however, some demonstrations of electro-holography using liquid crystal devices were presented. In this paper, we describe the process, characteristics, and driving methods of a liquid crystal device that aims for a fine pixel structure. Then, we discuss the design of an electro-holography system for the viewpoint of the spatial frequency of the hologram. We also experimentally examine the influence of the pixel structure on image quality by comparing electro-holography and optical holography images.
This paper presents a fast, new image-based technique for determining the connectedness of iterated function system attractors. It explores the use of a iterated function system (IFS) as a geometric representation for shape interpolation. For each shape interpolation, a two parameter family of iterated function system is defined, and a connectedness locus for these shapes is plotted, in order to maintain connectedness during the interpolation. The algorithm generates a smooth continuous shape transformation sequence from two given IFS attractors, initial and final, using polar decomposition. Polar decomposition is used for interpolation because its factors are unique, coordinate independent, simple and efficient to compute, and have physical and visual interpretation which are not found in other decomposition methods. Moreover, we apply it for natural image animation because it extracts the individual transformation components from a general affine transformation matrix and provides a basis for more controlled shape interpolation.
Chromakey method is one of key technologies for realizing virtual studio, and the blue portions of a captured image in virtual studio, are replaced with a computer generated or real image. The replaced image must be changed according to the camera parameter of studio for natural merging with the non-blue portions of a captured image. This paper proposes a novel method to extract camera parameters using the recognition of pentagonal patterns that are painted on a blue screen. We extract corresponding points between a blue screen and a captured image using the projective invariant features of a pentagon. Then, calculate camera parameters using corresponding points by the modification of Tsai's method. Experimental results indicate that the proposed method is more accurate compared to conventional method and can process about twelve frames of video per a second in Pentium-MMX processor with CPU clock of 166MHz.
A rangefinder system which measures both the 3-D shape and color texture of a human face is described. In our system, high-speed face measurement is achieved by using a space-encoding method with laser scanning. This system equips a color CCD camera as an image input device which grabs both range and color texture images. There is the possibility that laser radiation is harmful to human eyes and our system meets the Japanese radiation safety standards for laser products. In our experiment, this system obtained 3-D shape and color texture information in about 2 seconds. The measurement error on shape was found to be less than 1 mm.
MPEG-4 has information on object contours and is expected to be used for various applications. This paper tries to use the contour information for 3-D motion estimation. First, this paper describes a method of estimating that 2-D motion vectors of an object contour by matching feature points of the object contour. Then it describes the 3-D motion estimation method based on regulation theory by combining contour motion and the motions of several points in the object. Experimental results show that our method gives better performance compared with the conventional method.
In the conventional gradient method of optical flow the movement of an object is estimated by a bottom-up algorithm starting from local processing. However, the method cannot be applied to the large displacement of the object. Many methods have been proposed to cope with this limitation. As one such method, we have proposed an optimally scaled multigrid optical flow algorithm in which the expanding flow scale between consecutive multigrid levels is optimized. We have also proposed an orthogonal functional expansion method as a novel optical flow method. The orthogonal functional expansion method is a top-down method where the total image displacements are expanded from a low-frequency term to that of a high-frequency term. This method is expected to be applicable to the flow estimation with large displacements and deformation with expansion or shrinkage, which are difficult for conventional optical flow methods to cope with. Here, the apparent displacement field is calculated iteratively by a projection method which utilizes derivatives with regard to the invariant constraint of the brightness integral value. In this paper, we apply our proposed method to several real images in which the objects are displaced largely through rotation or deformed with expansion. We demonstrate the effectiveness of the proposed orthogonal functional expansion method by comparing it with conventional methods including an optimally scaled multigrid optical flow algorithm proposed by us.
The induction field model can be used to explain some experimental psychological results of a pattern recognition process for ambiguously shaped patterns. A lot of psychological experiments have been carried out to find the recognition process and the features of ambiguously shaped patterns. However, there is no mathematical model that can explain the results of psychological experiments for ambiguously shaped patterns. Recently, the focus has been on dynamics as the concept that explains the recognition process. Furthermore, the induction field model based on dynamics can explain some psychological experimental results for some artificial patterns and letters. This paper verifies that this model can explain some psychological experimental results for ambiguously shaped patterns.
We have developed a 2-D/3-D compatible display without special glasses. We set the image-splitter between the LC panel and the backlight in order to achieve the 3-D images without special glasses. We set a controllable scattering shutter between the LC panel and the backlight in order to display not only 3-D images without special glasses but also normal 2-D images without losing the resolution of the LC panel. By controlling both the image and the special shutter, a 3-D image and a 2-D image can be created in any rectangle groups of 16 rectangles on the screen. A 2-D image can also be viewed without losing resolution. Furthermore, this display has a 3-D indicator to make it easier to find the proper viewing area without the special glasses. We applied these technologies to a 3-D system on the Internet.