2018 Fiscal Year Final Research Report

The development of the new system for the automatic documentization of surgical operations

Research Project

PDF

Project/Area Number	16K12911
Research Category	Grant-in-Aid for Challenging Exploratory Research
Allocation Type	Multi-year Fund
Research Field	Medical systems
Research Institution	Kobe University
Principal Investigator	Harada Hitoshi 神戸大学, 医学研究科, 医学研究員 (70771517)
Co-Investigator(Kenkyū-buntansha)	宮尾祐介国立情報学研究所, コンテンツ科学研究系, 教授 (00343096) 金治新悟神戸大学, 医学部附属病院, 助教 (10637052) 掛地吉弘神戸大学, 医学研究科, 教授 (80284488)
Research Collaborator	SATO Yoshinobu Otake Yoshito Reiter Austin
Project Period (FY)	2016-04-01 – 2019-03-31
Keywords	画像 / 機械学習 / アルゴリズム / グラウンディング / 生体医用画像
Outline of Final Research Achievements	For the automatic documentization of surgical operations, we worked on developing the algorithm for recognition the various elements from the video of surgical operation, such as anatomical structure, surgical instruments, operative motion, and events during surgery. As a result, this algorithm has not reached the accuracy as we expected to the documentation of surgical operation. The difficulty in improving accuracy due to differences in anatomical structure and plasticity of organs among cases is cited as a factor of incomplete algorithm. However, it achieved relatively high accuracy for automatic recognition of less plastic objects such as surgical instruments.
Free Research Field	コンピュータ外科学
Academic Significance and Societal Importance of the Research Achievements	手術動画の自動認識アルゴリズムの開発において，動画より認識が容易な対象物と認識の困難な対象物の特徴を把握できたことはきわめて有用と考えられる．グラウンディング技術は現在多方面で発展が期待されている分野であり，外科領域を含む医療分野におけるグラウンディング技術の応用にとって重要な試金石となった．今後，精度の高い手術器具の自動認識を基に，手術動画の文章化のアプローチ法を検討することで，さらなる発展が期待できると考える．