Study of digital watermarking technique is making progress to protect intellectual property rights of digital contents. As a method of digital watermark, a method to add a watermark in the frequency domain is often used. In this method, only a few frequency components are changed according to watermark information, so that the pixel values sometimes exceed the dynamic range and watermark information is lost. This problem even occurs without attacks like image processing. In this paper, we propose two methods to prevent loss of watermark information, applicable to such cases, it is said that, one is the method called “inversion of increase and decrease” and another is “conservation of amounts of changes”. We can extract the watermark information correctly under the condition of no attacks by using these proposed methods.

View full abstract

Previous fragile watermarking algorithms have been kept in secret, and not been discussed about the security except²⁾. In 2), a secure fragile watermarking algorithm was proposed, even if it opens to the public. However, the algorithm needs to use different keys for different images. In this paper, we propose a fragile watermarking algorithm which can use the same key for different images. We show that this algorithm is secure, if it uses secure cryptographic function, pseudo-random number and self-synchronized pattern.

View full abstract

Stereoscopic displays without glasses are required in various fields. Stereoscopic display with back-light division system is one of those are proposed recently. This stereoscopic display emits light corresponding to position of observer's eyes by means of a large lens and shows parallax images, so this system is able to give observers stereoscopic images. But observers in different distances can't observe stereoscopic images at same time in this system, as the lens must move at a suitable position mechanically. In other words, when this system fixes the distance between the lens and back-light panel, the distance between the lens and observers is fixed also. In this paper, we suggest a new system to provide wide viewing area. To enlarge viewing area, the back-light panel is put slanting. It gives freedom of the distance between the lens and lights on the back-light panel. Because of observer-tracking program detects observer's iris on CCD camera images, recognizes observer's position by the distance between the eyes of the observer and emits light for position corresponding to observer's eyes, viewing area are also extended. Since lights on the back-light panel focus at the viewing points vertically, a diffusion filter is used to spread light vertically. So a few observers can view stereoscopic images even if they are in different distances at the same time.

View full abstract

The image template matching based on the normalized correlation coefficient is known as a general-purposed method because of its robustness of brightness change. However, the method requires large computational cost to match for all pixels included in the template image as well as in an objective image. This paper proposes a method for high-speed template matching that uses the side-edge pixels both in the template image and in an objective image. The side-edge pixels are located near the boundaries between a pair of planes that have different brightness. The experimental results show that the proposed method is robust to the movements of objects and to the brightness change in comparison with the template matching method based on the normal edges that are located just on the boundaries between planes. The processing time was reduced to 30% in comparison with the conventional method.

View full abstract

The staircase illusion is a visual phenomenon ever known in the field of psychology/psychophysics on binocular vision. We found two types of new staircase illusions, similar to this phenomenon, induced by a vertical square-wave gratings pattern tilted around horizontal/vertical axis. One is a round edged staircase illusion, the other is a vertical strip type staircase illusion. Then, we tried to apply our binocular vision model into these new phenomena to clarify and formulate the mechanism, and succeeded to visualize the whole shape of their 3D structures. Our computed and visualized results were very similar to and reproduces those ones observed by humans in the real world. As a result, we could confirm that computational simulation and visualization based on a binocular vision model are effective in the research of psychological visual phenomena such as the staircase illusion.

View full abstract

With popularity of Digital Still Camera-DSC, higher image quality is required. One of the subjects is that invisibility of shadow area caused by the narrow dynamic range of the CCD elements is improved automatically. Conventionally, gamma transformation, histogram equalization, and etc. have been utilized for the improvement but they were not always good enough. Recently, examinations applying to Retinex theory taking into account of human eyes characteristics proposed by Land are paid attention. This algorithm renders image clearly in shadow area effectively using spatial information between surrounding pixels arranged in two dimensions. However, some subjects such as dependency by adjustment parameters for output images and decrease of the density have been pointed out. This paper describes about the improvement algorithm which solved these subjects for contrast based on the Retinex theory and the effectiveness is shown by experiments using actual DSC images. And also, we describe about processing method of this algorithm at a higher speed.

View full abstract

Simulation of weather scenery in computer graphics (CG) is an important issue that has a wide range of applications related to image content, such as in movies and in various landscape simulations. In general, research to date into CG lightning has focused on generating a lightning pattern by applying probability and statistics techniques, which do not consider the generating mechanism of lightning. In natural lightning, an electric discharge path (stepped leader) is generated under the influence of an electric field formed by electrification of a cloud. In this paper, we present a method that generates a lightning pattern by simulating a stepped leader in an electric field defined by placement of electric charges in a cloud and ground.

View full abstract

In the environment of the SEM (Scanning Electron Microscope), it is necessary to establish the technology for recovering 3D shape of a target object from the observed 2D shading image. SEM has the function to rotate the object stand to some extent. This paper uses this principle and proposes a new method to recover the object shape using two shading images taken during the rotation. The proposed method uses the optimization of the energy function using Hopfield neural network, which is based on the standard regularization theory. It is also important to have the initial vector that is close to the true optimal solution vector. This paper introduces the method to give the suitable initial vector using observed two images during rotation. Computer simulation evaluates the essential ability of the proposed method. Further, the real experiments for the SEM images are also demonstrated and discussed.

View full abstract

There are several methods for image stabilization function of camcorders. This paper describes cost-saving technology concerning an electric image stabilization system which is one of these methods. In the electric system, an image buffer memory occupies large part of system hardware, and it is most effective to delete it for minimization of system cost. However, a sufficient stabilization effect is not obtained when the buffer memory is deleted, because it works as a compensator for delay time of 1 field period caused by motion detection processing. Therefore, compensation for the delay time utilizing another means (a motion prediction filter, for example) is required for improvement of the stabilization effect. However, improvement in all frequency range of image fluctuation is difficult. For the reasons mentioned above, we have evaluated human visual perception to make sure of specific frequency range that is an eyesore, and have designed a motion prediction filter to improve performance in that range. As a result, improvement of stabilization effect by using the motion prediction filter has been confirmed through subjective evaluation.

View full abstract