Diago L.,Interlocus Inc |
Diago L.,Meiji University |
Romero J.,Interlocus Inc |
Romero J.,Meiji University |
And 4 more authors.
Advances in Intelligent Systems and Computing | Year: 2017
Soft-computing forms the basis of a considerable amount of machine learning techniques which deals with imprecision, uncertainty, partial truth, and approximation to achieve practicability, robustness and low solution cost. This paper describes an application developed to understand what means a picture (portrait) to be Iyashi. The neuro-fuzzy quantification allowed extracting a set of 35 rules that describe the meaning of the word Iyashi to hundreds of users. Facial expressions of the subjects and their brain signals during the evaluation of the images have been explored to validate the obtained rules. The developed system allows discovering the rules that describe the preferences of users while exploring the space of possible design parameters so that the system predictions match the preferences of users. Interactive genetic algorithms (IGAs) have been used for the implementation of a color recommendation system following customer’s preferences. The combination of color and geometric shapes is also explored. © Springer International Publishing Switzerland 2017.
Li W.,Nanjing University of Information Science and Technology |
Wu Z.,Shanghai JiaoTong University |
Junichi S.,Interlocus Inc. |
Hagiwara I.,Tokyo Institute of Technology
Communications in Computer and Information Science | Year: 2012
Recently, some results of the G 1 continuity conditions for B-splines surfaces have been presented. These G 1 conditions can be used in the reconstruction of bicubic and biquintic smooth B-splines surfaces with a single interior knots. However, the C 1 continuity conditions of B-spline surfaces with arbitrary degrees have not been solved. In this paper, we obtain the C 1 continuity conditions between two adjacent B-spline surfaces with arbitrary degrees. We also present a practical scheme of reconstructing model using the C 1 continuity conditions in reverse engineering. © 2012 Springer-Verlag.
Zhang Z.X.,Tokyo Institute of Technology |
Hagiwara I.,Tokyo Institute of Technology |
Savchenko M.,Tokyo Institute of Technology |
Feng Y.X.,Tokyo Institute of Technology |
Shinoda J.,Interlocus Inc.
Advanced Materials Research | Year: 2011
In this paper, a robust tetrahedral mesh generation method based on Advancing Front technique is proposed. The proposed method inherits advantages of Delaunay method and Advancing Front method, such as efficiency of Delaunay method and maintaining the given boundary triangle mesh exactly of advancing front method. Tetrahedral mesh is generated from the given triangle surface mesh. This method mainly includes three stages. Firstly, the minimum container box of the triangular surface mesh is calculated and points are inserted into the box. Then the proper point is selected out to generate tetrahedron's layers from surface to the interior volume of the model, so g the surface mesh can be maintained. The operation is simplified, and calculation efficiency is also higher than common Advancing Front method. At last, triangle intersection is examined. This technique allows generating the tetrahedral mesh with high quality elements with surface mesh preservation. A shoes model with both convex and concave surface is chosen for the experiment. The result clarified the robust and high efficiency of the proposed algorithm. © (2011) Trans Tech Publications.
Rodriguez L.,Interlocus Inc. |
Diago L.,Interlocus Inc. |
Hagiwara I.,Meiji Institute of Advanced Studies of Mathematical science
Journal of the Institute of Image Electronics Engineers of Japan | Year: 2012
An efficient image segmentation is important for computer vision in any attempt to analyze or interpret an image automatically. The automatic segmentation based on color in a natural image is very hard. So, semi-automatic segmentation methods incorporating user interactions are becoming more and more popular. Our approach makes possible that the image can be separated in many regions starting from a mean shift over-segmentation followed by the user interactions. After the over-segmentation, the user selects the regions of interest and makes the interactive operations. In order to avoid the fatigue in the user and to decrease the user interactions, three operations are made semi-automatically using the polygonal representation of the regions. The polygons are constructed using a new points connection algorithm marking branch points and removing overlapped connections. The proposed interactive image segmentation method is compared with well-known related algorithms in terms of the accuracy and efficiency. Although the accuracy of the proposed segmentation result depends on the user interactions, our method increases the object and boundary accuracy in 8% and 20% respectively in comparison with graph cuts methods using less time. Compared with an algorithm based on the results of mean-shift segmentation, proposed approach needs more user interactions but the segmentation results are more accurate according to Jaccard index and visual information, mainly in the boundary part.
Zhang Z.,Tokyo Institute of Technology |
Savchenko M.,Tokyo Institute of Technology |
Feng Y.,Tokyo Institute of Technology |
Fukuhisa T.,Interlocus Inc. |
And 2 more authors.
Transactions of the Japan Society for Computational Engineering and Science | Year: 2011
Segmentation is a critical and necessary procedure of reverse engineering with the aim of partitioning scanner data or polygonal meshes into meaningful parts. In this paper, an automatic segmentation technique is proposed for triangular meshes with the identification of surface types of the mesh segments according to the following surface's classification: the planar, cylindrical, conical, spherical, and toroidal surfaces, extrusion surfaces, ruled surfaces, and filleted surfaces. Sharp edges are detected and identified as boundaries of segments. An assumption-verification method is developed for extracting the mesh segments which identified according to surface's classification. The sequence of the extraction of the mesh segments is determined according to the dimensions and shapes of the Gauss maps of the segments. The algorithm starts from calculating normal vectors and principle curvatures of the neighborhoods of nearby mesh vertices of the triangle meshes to identify sharp edges and areas with high curvature. Triangles of the segments are assembled based on region growing method where the sharp edges are defined as boundaries of the segments. A fundamental partitioning the surface mesh based on sharp edges allows the accurate identification of the surfaces of the segments based on delicately proposed fitting technique and statistical error distribution. © 2011 by the Japan Society for Computational Engineering and Science.