2018 Fiscal Year Final Research Report
Exploring image representation that enables to detect objects with little visible features and its application for protozoa identification
Project/Area Number |
17K20025
|
Research Category |
Grant-in-Aid for Challenging Research (Exploratory)
|
Allocation Type | Multi-year Fund |
Research Field |
Applied informatics and related fields
|
Research Institution | Japan Advanced Institute of Science and Technology |
Principal Investigator |
Yoshitaka Atsuo 北陸先端科学技術大学院大学, 先端科学技術研究科, 准教授 (60263729)
|
Co-Investigator(Kenkyū-buntansha) |
所 正治 金沢大学, 先進予防医学研究センター, 准教授 (30338024)
|
Research Collaborator |
Muhamad Kamal Mohammed Amin
Pho Ngoc Dang Khoa
|
Project Period (FY) |
2017-06-30 – 2019-03-31
|
Keywords | 寄生虫 / 原虫 / 物体検出 / 種の識別 / 局所特徴 / 分子疫学 |
Outline of Final Research Achievements |
Protozoa that is classified into parasite causing diseases can be identified not only by DNA analysis but also by its morphological features that is observed with microscope. In order to improve the accuracy of identifying species of parasite, we proposed to represent visible feature of a parasite with local features of different scale divisions. Detection of the region of parasite in micrograph is carried out by applying original deep neural network. Moreover, parasite database is enriched by epidemiological study for clarifying differences of related species.
|
Free Research Field |
画像処理
|
Academic Significance and Societal Importance of the Research Achievements |
寄生虫は顕微鏡のみで確定診断が可能な感染症の病原体である。寄生虫感染症は国内症例が極めて少なく、顕微鏡検査で寄生虫を検出した経験のある検査技師が少ないため、習熟度の低下が問題となっている。本研究では、種ごとに固有な視覚的特徴の差が大きくない原虫などの寄生虫をより精度高く識別する特徴量表現により虫体を識別する手法を提案し、また、寄生虫情報データベースの拡充により上記課題に対する情報提供システムの基盤を提供する意義がある。
|