Ayumu Saitoh

On selection of intraocular power formula based on data classification using self-organizing maps

In this paper, a method of selecting one type out of three types of intraocular lens (IOL) power formulas for data of each individual cataract patient is presented, using self-organizing maps (SOMs). The proposed method employs three-dimensional vectors each of which has measured values of axial length, corneal refractive power and cylinder as training data for maps. The first two values are substituted into the power formulas, while the last value is associated with the astigmatism. This paper also proposes neuron labeling that depends on postoperative refractive errors occurring under the assumption that each of the three power formulas is applied. The proposed method determines the formula to be applied to some patient, observing the label attached to the winner neuron for the presented data with the above three element values associated with the patient. The experimental results finally establish that the proposed method adequately works to select the formula.

Development of Dual‐Reciprocal Extended Boundary‐Node Method

The dual‐reciprocal boundary‐node method has been reformulated without using integration cells and its performance has been investigated by comparing with the dual‐reciprocal boundary‐element method (DRM). The results of computations show that the accuracy of the proposed method is higher than that of the DRM. In addition, when the number of boundary nodes exceeds a certain number, it is found that the speed of the proposed method is almost equal to that of the DRM.

Video object segmentation using color-component-selectable learning for self-organizing maps

In this article, self-organizing-map-based video object segmentation is proposed, assuming that either Y-quantification or HSV-quantification can be systematically selected. Given a video sequence, the value of the probability density function for each component value is calculated according to a kernel estimation at the first frame. Some areas randomly chosen from the background are then examined, using each component value, to judge whether or not they include the target object. The quantification is determined so that the frequency of occurrence of false extractions can be reduced. The data presented to the maps are generated based on the selected quantification. Experimental results show that the proposed method recognizes the target object well.

Detection of tumors on stomach wall in X-ray images

Double contrast (DC) X-ray images are useful and cost-effective for the diagnosis of stomach tumors. However, it is difficult to automatically extract tumors from these images. This is due to the variations in the tumors appearing in the images as a result of the changes in stomach shapes and the distribution of barium meal in the stomach. In this paper, we propose an automated method for detecting tumors in DC X-ray images. Our method utilizes the distributions of contrast gradients in the images. The performance of our method is demonstrated by using actual DC X-ray images.

Development of three-dimensional extended boundary-node method for potential problem

The three-dimensional boundary-node method has been reformulated without using any integration cells. For evaluation of the influence coefficients, an implicit surface is assumed as a surface boundary and the 3D polar coordinate is employed. Numerical experiments show that, for the Laplace problem, the accuracy of the proposed method is almost the same as that of the boundary-element method with linear elements.

Development of two-dimensional meshless approaches without using integration cells

The Element-Free Galerkin Method (EFGM) and the Boundary-Node Method (BNM) have been reformulated without integration cells. After a boundary is represented in terms of an implicit function, matrix elements are evaluated by use of the function. The results of computations show that the accuracy of the reformulated BNM is even higher than that of the dual-reciprocal boundary-element method.

New Method for Implementing Essential Boundary Condition to Element‐Free Galerkin Method

The new method for implementing the essential boundary condition to the Element‐Free Galerkin Method (EFGM) have been proposed and its performance have been evaluated as compared with the classical EFGM in the Poisson problem with the Dirichlet condition. The results of computations show that the numerical solution of the classical EFGM cannot satisfy the essential boundary condition whereas that of the proposed method satisfies the essential boundary condition exactly. Therefore, it is found that the proposed method is more suitable than the classical EFGM for solving the Poisson problem with the Dirichlet condition.

Generation of a Smooth Implicit Function Distributed Over the Entire Domain from Large Scattered Point Data

For the purpose of generating a smooth implicit function distributed over the entire domain, a method for smooth combination of two implicit functions f(x) and g(x) has been proposed. The method can be applied to large scattered point data, since the implicit function f(x) is generated by the partition of unity based method. The other implicit function g(x) is generated by the radial basis functions based method with a global support basis. For smooth combination of f(x) and g(x), an appropriate weight function is employed. Numerical experiments demonstrate that the proposed method enables to generate a smooth implicit function F(x) distributed over the entire domain. In addition, on the given points, accuracy of f(x) is exactly the same as that of f(x). Furthermore, the computational cost for generation of f(x) is almost the same as that of f (x).