In recent and future informationalized society, it is needed that a scriber can choose and get the desiered information from various information quickly service for customer. So, for best we developed Advanced Information Supply System on a Broadband Network which can offer the information concerned with our business and many other types of information by requests. This system also have a notice service. This system is based on Videotex using NAPLPS and offer the information by still pictures and narration.
In this paper, we propose an optical 3-D range-finder for a moving robot, an unmanned carrier,etc. This is based on a spatial-filtering using for an optical system such as an infrared seeker. In this system, some spatial-filters are arranged at equal intervals in an optical system's image-plane because of getting the azimuth's information of obstacles. We confirmed availability of the system through an experiment by using a CCD camera for digital-image processing.
In this paper, the new coding scheme for binary images which is called PCSB (Rogressive Coding Scheme using Block Reduction) is proposed;(1) It has the following characteristics; R It has the functionality of progressive build-up as well as sequential buildup. (2) The new image reduction method is applied. Then, the image quality of intermediate stage is better than .that of conventional methods. (3) It is applicable to not only images of line drawing and printing type but also the half-tone images which are represented by the dither method. (4) The coding efficiency of the proposed method is 16% better than that of MMR coding scheme.
A T R Auditory and Visual Perception Research Laboratories ABSTRACT Image data compression techniques are fundamentally based on the suppression of redundancy, maintaining an acceptable image quality. In this paper, the applicability of Neural Networks to image coding is discussed. We proposed a local-parallel type multilayer Neural Network model which can get abilities of image data compression and reconstruction by Back-propagation learning rule. Computer simulation experimentally shows that image compression and reconstruction ability can be realized on Neural Network for natual facial images.
Human face model using free-form surfaces has been deveioped. A shape of this model is generated by letting free-form surfaces pass thgrough the feature points. The basic shape of a face is made by inputting the feature points of an expressionless fasicial images that are photographed from front and side. And facial expressions can be changed by moving the feature points adequately to transform surfaces. By using this method and mapping the facial images to this shape, realistic animations can be produced easily for various faces.