Recently, a new image compression method based on a fractal theory was proposed. It assumes that image redundancy can be efficiency exploited through self-transformability on a blockwise basis. And it approximates an original image by a fractal image. In this paper, we evaluate this compression technique and propose some new applications based on it.
We have developed a transmitting system using a satellite for intelligent communication of sign language images. This paper describes the configuration of the system, and the description of the arm's motion in sign language dictionary for image synthesis.
In this paper, we propose the virtual studio which is an expansion of the electronic studio set technique which has been developed in the field of image production. In the virtual studio, the virtual camera shoots the images of actors taken in advance by actual TV cameras at the different space and time, which are located in the electronically generated studio set. Producer in the virtual studio controls the whole system with three man-machine interfaces for a camerawork, a change location of the actors and a modification of the set. The proposed system, imaging technique using the system and the future of the virtual studio are discussed.
This paper proposes a scene describing method for video production. The method describes contents of a scene using Video Components and Script. The Video Components are corresponded to images of scene constructing objects in the real world. They include not only image data but also attributes of the scene constructing objects and capturing conditions of images. The Script describes the relation among the Video Components and the attributes of the scene. We also discussed an application of the method for video browsing.
This report proposes a method to easily span hypermedia links between a specified object in a video and its related information. It automatically extracts video portions in which a specified object appears. And it calculates image regions of the object in each frame of those portions. These regions are used for nodes of the hypermedia links. The portion extraction algorithm in this method is also used for giving keywords for video databases. Keywords given for one scene, concerening with a certain object, can be automatically given to other scenes which contains the video portions in which the same object appears.
An efficient coding method for octree representation of 3-D objects is proposed. In this method, a Marcov model referring neighbor nodes on upper layer is used to reduce the entropy. A technique to reduce the number of contests is introduced as well. Experimental results reported here indicate that sample data can be compressed about 40% better then conventional methods.
Both the brightness and contrast ratio of an eye contact type display apparatus are measured under various conditions. A projection screen that is controllable In both the transparent and translucent state is used for eye contact. Simulated characteristics based on the model proposed here are compared with the experimental results. The optimum display condition is when both the screen driving period and camera shutter time are short as possible when the light gain is large.