著者

高津 和紀 高田 宗樹 平田 隆幸

雑誌

福井大学大学院工学研究科研究報告 (ISSN:18821871)

巻号頁・発行日

vol.67, pp.1-10, 2018-10

著者

高津 和紀 高田 宗樹 平田 隆幸

出版者

福井大学大学院工学研究科

雑誌

福井大学大学院工学研究科研究報告 (ISSN:2433815X)

巻号頁・発行日

no.68, pp.1-11, 2020-03

---

Sota Fujii is a professional shogi player who has achieved renewal of historical records. As Fujii is a young man, his brain is in the growing stage. Therefore his ability of shogi is improving with his brain growth. Analysis using shogi AI characterizes the future of the player's shogi quantitatively. There is a possibility that we can detect the rapid growth of the young player by analyzing only a few games. In this study, Fujii's ability was evaluated by analysis of records of shogi using multiple shogi AI.

著者

井上 敬章 平田 隆幸

雑誌

福井大学工学部研究報告

巻号頁・発行日

vol.53, no.2, pp.127-137, 2005-09-30

---

Quantitative characterizations of gray scale images of typical patterns ware carried out by calculating, information entropies. We calculated the information entropy (H_<gs>) of the coarse grainined gray scale patterns by changing the size of coarse graining(s). Typical patterns were constructed by changing both a basic figure and its spatial distributions. Our attempt is to characterize the gray scale images by the graph of Hgs vs. s. We demonstrated that the shape of the graph of H_<gs> vs. s could distinguish among typical patterns.

著者

平田 隆幸 藤原 郁弥 藤本 研治 園山 輝幸 原田 烈光

出版者

福井大学

雑誌

福井大学工学部研究報告 (ISSN:04298373)

巻号頁・発行日

vol.51, no.2, pp.215-220, 2003-09-30

被引用文献数

3

---

Yes. The principle of a kaleidoscope can magnify tiny differences of dot patterns. Human brains can recognize a small difference of patterns especially in the case of patterns having some symmetry. The kaleidoscope generates 3-fold rotation symmetry patterns by using the mirrors so that we can recognize tiny differences of time-series by embedding the oeak-value on the scope of the kaleidoscope. We demonstrate this effect using synthetic data and then apply this method to RF echo signal of ultrasonic diagnostic system.

著者

平田 隆幸 黒岩 丈介 浅井 竜哉

出版者

福井大学

雑誌

福井大学工学部研究報告 (ISSN:04298373)

巻号頁・発行日

vol.52, no.2, pp.161-167, 2004-09-30

---

A numerical simulation of Hodgkin-Huxley model was carried out by using a Runge-Kutta method. In the numerical simulation, the functions of Numerical Recipes in C were used for solving the Hodgkin-Huxley equation. The accuracy of numerical solutions was discussed for both simple Runge-Kutta method and adaptive stepsize control Runge-Kutta method. The difference between simulation performed by using float type variables and one by using double type variables was also discussed. A large neural network of Hodgkin-Huxley neurons was carried out. A synchronization in the neural network was observed by changing the weight of synaptic transmission.