We have developed a new gaze analyzer which can limit the viewer's gaze time by inserting a masking pattern in his focusing area. The size of the masking pattern can be set arbitrarily. Moreover, the time to start masking and the time to retrieve the original image can be selected arbitrarily. This system has various convenient functions for executing psychological experiments, such as generation of the presentation patterns, their storage on optical disk, automatic editing of the experimental procedure and its execution, etc.. Details of this system are described and the examples of the visual search experiments are introduced.
Binocular eye movements are measured while showing a moving stimulus to one eye and a still stimulus to the other. To perform the experiment, we constructed a new 3D image processing system and a binocular eye movement analyzer. This system can control the 3D image output, measuring binocular eye movements simultaneously. Analyzing the measured binocular eye movements, we found that conjugate eye movements frequently occured in comparison with vergence movements, and that the dominant eye had a tendency to achieve peak velocity eariller than the other eye even if when the dominat eye didn't see the moving stimulus.
We quantified the degree of coherence of small involuntary (eye) movements between the left and right eyes on the vertical and horizontal components, respectively, as a chaotic phenomenon. We measured the drifts of binocular small involuntary movements with independent vertical and horizontal components using the limbus reflection method. After extracting the drift components from the measured data, we calculated the fractal dimension of binocular small involuntary movements on the vertical and horizontal components, respectively. Then we found that there is a difference between vertical and horizontal binocular components in higher dimensional data space. This result reflects the qualitative phase difference between horizontal and vertical drift components of small involuntary movements.
A procedure to transform any given closed curve to an equilateral polygon approximation is proposed. In the procedure an iterative algorithm proposed for obtaining equilateral polygons and Uesaka's P-type descriptor play important roles. The condition for the convergence of the iterative procedure is investigated. Experimental results are also shown to demonstrate that the proposed procedure works well.
We propose a method to detect circles in digital images. The method uses the geometrical property that the perpendicular bisectors of the segments joining two points on a circle pass through the center of the circle. This method is less sensitive to noise and digitizing error than the previous method using the gradient.
A digital image processing technique for measuring velocity and size distribution of moving particles is proposed. The technique is based on the theory of spatial filtering velocimetry and that of light scattering. The usefulness of the method is confirmed by utilizing artificial image sequences created by computer simulation.
The Hi-Vision receiver, together with its realism-packed large-screen display and stereo sound, is expected to play a central role as a video information device in every household of the future. And while the Hi-Vision telecasts via satellite are now being broadcast in the analog format, it is quite certain that the Hi-Vision signal source will likewise make the transition to the digital format as the digitalization of the various related media progresses. To this end, we suggest a digital interface for the MUSE decoder. We have developed a special LSI to control the decoder and have produced a test version of a MUSE decoder equipped with the digital interface. We also put together a Hi-Vision audio-visual system and confirmed the various features and functions. This report will cover these developments.
New concept of intelligent handset for visual communication is described in this paper. Fundamental requirements and basic functions are discussed. Opinion scores are tested concerning facial expressions with partial artificial coverings. Based on these evaluations, prototype system is realized.
CCITT has standardized the specification of video codec for audiovisual services at p*64Kbps as Recommendation H. 261. Hence the video codec based on H. 261 has the compatibility among different vendors. We have implemented the VLSI chip set for motion picture decoder based on H. 261. Because the chip set is composed of only three processors, it is possible to realize a compact and cost-effective system. Furthermore we have implemented the multimedia system which is an application of the chip set.
This paper describes the design and the implementation of a visual multi window system for workstations and personal computers and also describes system configurations of a prototype system. This window system can display both graphics and 11 full motion video pictures on a same display device simultaneously with Visual Multi Window control module. User can also handle a full motion video window with same user interface as an ordinal graphic window on workstation environments.
Many high performance personal computers and workstations have been developed. Especially, recent models of high performance workstation have got similar performance of mini-computers which were sold in several years before and they are used not only at science fields but also at business fields. Some of them are able to handle multimedia, for example video, audio, images and hand-written images. These media, however, are used independently and not unified at the point of the graphical user interface. This paper describes about multimedia graphical user interfaces. We studied multimedia windows and icons. These multimedia are unified using multiwindows and icons. We made the prototype personal multimedia multi-points teleconferencing terminal (PMTC) using multimedia windows.
In the diagnosis of mental disease such as Alzheimer's, quantative evalution of brain atrophy ratio (b.a.r) is quite important, and it is quite natural to use commonly distributed X-ray CT image for this objective. However, in X-ray CT image processing, a fundamental problem to be resolved is how recognize vague boundary of each brain subarea, such as cerebral, cerebral ventricle and cerebral spinal fruid area. To resulve this problem we adopted a new method based on analizing each CT image data statistically. The detail of the method and its application results is discrived in this paper.