In the society of the 21st century, ensuring a comfortable life where one is able to enjoy the benefits of advanced computer equipment and social systems without regard to race, sex, occupation or level of education is an important issue. Gifu Prefecture places much importance on this issue and believes that the market for this field will expand in the future. To this end, it is crucial to research about how to make a more advanced interface for humans and computers. This interface needs to address issues such as how computers can correctly understand humans and how computers can correctly relay information to humans. To accomplish this, we work with computer equipment using image processing technologies. Based on these technologies, we are able to give equipment the ability to sense, judge and recognize humans and their surroundings.
We have developed a new type of virtual studio in which a real space image and virtual space image combined naturally with no boundary seam. When adopting this new virtual-real hybrid system, there are two major advantages. One is that the actors can concentrate on their role in the real studio sets, and second is that camera work can be done without worrying about the view outside the studio sets. We constructed an omnidirectional image with ultra high-definition and combined it as a virtual studio image with a real studio image. It was shown that smooth and natural merging of the omnidirectional images and the real studio images can be done by adjusting appropriate the location of the principal point of a camera lens.
This paper proposes a method to measure the inclined angle of an objective image based on the most frequent edge direction. This paper also describes a relation between the measured inclined angles by the proposal method and the visual inclined angles. An edge direction is changed according to the inclination of an objective image. The most frequent edge direction is supposed to show the global inclination of the image. The proposed method measures the inclined angle of the image by detecting the edge direction with the peak in the histogram that represents the distribution of all edge directions in the image. The experimental results show that the measured inclined angles by the proposed method were similar to the visual inclined angles evaluated by the subjects.
This paper describes Keitaide-Music (“Music by Mobile Device”) framework. It is to manage users’ access to distributed digital data using the concept of “Super Distribution”, or flexible and legitimate peer-to-peer distribution, as well as Public Key Infrastructure (PKI). Keitaide-Music has also the scheme of content synchronization. It is something that audio can be played back with synchronizing image and text. This synchronization scheme is designed to be simple and suitable for mobile devices with very low power requirements. In April 2002, Sound Market introduced synchronized contents such as English conversation lesson.
MacAdam plotted the results on the chromaticity diagram in terms of the standard deviation of color matching in several directions about selected colors. The resulting figures formed closed curves on the diagram. The problem corresponds to a model fitting problem. After the MacAdam ellipsoid, model fitting criterions based on information theory such as the MDL has been formulated. This paper investigated the experimental results of the MacAdam Ellipsoid by using the MDL criterion.
A method to improve a contrast of a color image using a genetic algorithm is proposed. The accumulated edge strength in R, G and B channel image that compose an output color image is used for evaluating an individual fitness. The difference of color hues between on original color image and the output image is also used for evaluating the fitness. The experimental results show that the output images by the proposed method had good contrast and visual clearness compared with the results by the liner brightness transformation.
Color image data from an image acquisition device are transformed to colorimetric values to achieve device independent color reproduction at an early stage in a color management system. This transformation is usually called colorimetric calibration or color correction. The accuracy of the calibration depends on its spectral sensitivities of a set of color sensors and noise present in a device. In this article, a measure to evaluate the influence of noise on the accuracy of calibration is described and a method is discussed to predict the influence using prior knowledge of spectral properties of sensor sensitivities, reflectances of objects and illuminants is available.