檜垣, 泰彦

HIGAKI, Yasuhiko

光学的文字認識

自動マスキング

デジタル・スカラシップ開発

Plant Specimen Images

Optical Character Recognition (OCR)

Automatic Masking

Digital Scholarship Development

[Abstract] The Haginiwa plant specimen images, are one of the largest collection of flowering plants in Japan, are expected to be published at Chiba University Academic Resource Collections (c-arc) that adopts IIIF and Right Statements. However, there is concern about resource conservation that endangered species are recorded in the collection, and detailed collection sites are described. In this study, we focused on optical character recognition (Optical Character Recognition), which is a method for automatically detecting and masking the collection area described in the sample image. W e mainly use “Cloud Vision API” for OCR, and refined parameters including images itself. As a result of study, the masking area determination algorithm needs some adjustment, but we obtained the prospect of performing masking automatically.