We restore a blurred image caused by defocus of a lens using the shift-invariant wavelet transform created using RI-Spline wavelets.In a defocus blurred image,the blurring kernel varies depending on its position in the image,so a positional frequency representation,such as a wavelet space,is required in order to deblur it.To restore the defocus blur,we assume that the blurring kernel can be obtained regardless of its position in an image.In experiments we conducted using synthesized images,our method using the shift-invariant wavelet transform showed better deblurring performance than that of the ordinary wavelet transform.We also show that our method can be applied to real images if additional range data is used.

抄録全体を表示

Road signs provide important information to guide and regulate the behavior of drivers and pedestrians in order to make their journeys safer and more comfortable.However,the shape features and color information displayed on circular road signs are not always the same.We propose a method for recognizing road signs on the basis of the shape features and color information derived from a color scene image.The proposed method works in three steps.First,a process already proposed in one of our previous studies detects road sign areas' color features and circularity using fuzzy reasoning.Second,the non-standard shape features of the detected road sign area are determined.Third,the values obtained from every sub-area of the image are used to enable the recognition of road signs.Our experimental results suggest that the proposed method can accurately recognize different types of circular road signs.

抄録全体を表示

We propose a studio system called an “interactive virtual studio”,in which an actor looks directly at virtual objects and manipulates them in a virtual studio.In a conventional virtual studio,such as that often used for recent TV programs,real images of actors are taken by cameras in a studio and superimposed with three dimensional (3D) virtual objects created by computer graphics (CG),so that they are presented to the viewers.However,the actors in the studio cannot see the virtual objects nor manipulate them directly.We introduced augmented reality (AR) to the virtual studio in order to enable them to do this. In our interactive virtual studio,the actors wear wearable displays (WD) so they can see the virtual objects.The actors are thus able to manipulate the virtual objects directly by using a pointer with a three-dimentional sensor on its end.To make this system,it is necessary to maintain a spatial relationship among the different coordinate systems that describe the real space,the virtual space,and the video images presented to the viewer.In the previous work,this geometric registration was made by setting several sample points and collecting errors between the coordinate systems.In contrast,we carry out the geometric registration adaptively with the following process.The actor manipulates virtual objects based on the constraints between three positions:that of the virtual objects,that of the pointer used for manipulating the objects in the virtual space and that of the image presented to the viewers.

抄録全体を表示

This paper describes an evolutional long delay wave measuring apparatus for measuring delay profiles and levels accurately in the single frequency network (SFN) used for terrestrial digital TV broadcasting.Measurable delay times using the conventional scattered pilot (SP) signal method are theoretically restricted to less than 0.5ms.This equipment has no limit to the length of delay times it can measure. We fabricated and tested a proto type and confirmed that it can be used to measure delay profiles with delay times longer than 1ms.In addition, we discuss errors that occurred when using this method and the countermeasures that were taken.Performance was satisfactory. In the future, this apparatus will be effective for achieving coverage with excellent quality following the construction of a new relay station that will expand SFN. In addition,it will be effective for electric field strength measurements taken during maintenance and it will assist in the efficient operation of SFN by its ability to confirm the quality of coverage.

抄録全体を表示

We researched and developed a new system for score-following of orchestra performances that uses audio-signal amplitude.In the existing score-following system,pitches extracted from the sounds produced during the performance using spectral analysis.These sounds are matched with the pitches notated in the score.However,in orchestra music,many instruments are often played at the same time and the fundamental tones and overtones of these instruments overlap,making pitch-detection very difficult.However,the envelope of the amplitude of the performance forms distinct peaks at the note-onset positions.Therefore,we proposed and implemented a real-time score-following system that matches the amplitude peaks extracted from the performance sounds with the amplitude peaks that can be predicted by reading the score.

抄録全体を表示

We present improved practical color filter array interpolation methods that use color correlation for use in digital cameras.Cost-effective digital cameras use a single-chip image sensor and subsequent interpolation processing to produce full-color images.Therefore,many type of sophisticated algorithms have been proposed to create high quality images.Their complicated algorithms make it difficult,however,for them to be implemented in digital cameras due to limited computing capacity or available processing time.Although existing interpolation methods that use a directional color correlation are very cost-effective,they result in color artifact on the edge of color boundaries.We propose two algorithms to resolve this problem.One is a constrained color correlation method that is achieved by adding constrained conditions.The other is the precious color correlation methods that uses a nonlinear filter.The proposed methods can eliminate the artifact regardless of any color correlation of the images.Experimental results confirm the effectiveness of these methods.

抄録全体を表示

We present improved practical color filter array interpolation methods that use color correlation for use in digital cameras.Cost-effective digital cameras use a single-chip image sensor and subsequent interpolation processing to produce full-color images.Therefore,many type of sophisticated algorithms have been proposed to create high quality images.Their complicated algorithms make it difficult,however,for them to be implemented in digital cameras due to limited computing capacity or available processing time.Although existing interpolation methods that use a directional color correlation are very cost-effective,they result in color artifact on the edge of color boundaries.We propose two algorithms to resolve this problem.One is a constrained color correlation method that is achieved by adding constrained conditions.The other is the precious color correlation methods that uses a nonlinear filter.The proposed methods can eliminate the artifact regardless of any color correlation of the images.Experimental results confirm the effectiveness of these methods.

抄録全体を表示

We propose a simple method to generate a high-resolution (HR) image from a single low-resolution (LR) image.The method solves a simultaneous linear equation that relates the HR image to the LR image.Since this equation is underdetermined,we utilize the energy compaction property of discrete cosine transformation to reduce the number of unknowns used to represent the HR image.We propose two different implementation methods.The first divides the LR image into a set of small blocks and the proposed method is applied to each block separately.The second implementation method applies the proposed method directly to the full image frame.The results of our simulation studies are shown.

抄録全体を表示

As a photometer measures only a small region in the projected area,it is not suitable for measuring a large-scale display,such as the whole area produced by a projector.Instead of the photometer,the technique used in this research employs raw digital images,which have been recently noticed.Using raw digital images,we accurately measure the luminance distribution of the whole projected area for every pixel and obtain a transfer function of the luminance response for the projector.The transfer function is essential for luminance correction.This paper shows how using raw digital images is a key technique for luminance correction.

抄録全体を表示