Several kinds of studies on color image analysis have been carried out, such as automatic key words extraction from color images and automatic classification. of color image sequence. But, the recognition level of color images is not enough high to recognize all things from color images alone. To solve this problem, we are developping a multimedia cooperative drama scene recognition system using common concepts. For these efficient handling of color moving media, the recognition of human is indispensable. In this paper, a human extraction method using two step human model is shown with an experimental result.
This paper proposes a camera operation extracting method that uses spatiotemporal image to "index" video sequences. The proposed method consists of two stages, one is a video motion visualizing stage, and another is its extracting stage. In the visualizing stage, we define Video Tomography Method(VTM), and in extracting stage, a camera operating information is automatically extracted with video. To visualize. indexes and interactively control video sequences, we propose a new human-video interface. For a human-video interface, we enhance video icons. Experiments confirm the practicality of the proposed method and the enhanced icons.
This report describes a method for extracting a hierarchical structure of video sequences in time domain as a means for video browsing. First, a method for describing video content is proposed. Next, a hierarchical segmentation method based on the descriptive method is discussed and results of an experiment are also introduced. This experiment has proved effectiveness of the proposed methods.
it was evaluated that relations between quality of picture and the number of divided virtual space. Next, it was evaluated that relations between quality of picture and the rate of data compression of stereoscopic images. Those show us not to spoil the quality of stereoscopic images decreasing information with depth too much.
We have developed a transmitting system for intelligent communication of sign language images. This paper describes the making sign language images system and the method of controlling the speed of animation by meaning.
Now, Computer Graphics (CG) is tried to use in many fields, one of the most important problems is the means of modering, particularly, modering of nature object is important. This paper is about the means of modering of plant, especially popura. I successed in expressing the peculiar branch of popura by assuming virtual attractive force.