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.