Object Detection of Single Cells for Automated Single-Cell Encapsulation using LCD 3D Printer

カテゴリ: 研究会(論文単位)

論文No: BMS24016

グループ名: 【E】センサ・マイクロマシン部門 バイオ・マイクロシステム研究会

発行日: 2024/07/01

タイトル(英語): Object Detection of Single Cells for Automated Single-Cell Encapsulation using LCD 3D Printer

著者名: GHOUS Ammar(Toyohashi University of Technology),Hussain Chowdhury Rifat (Toyohashi University of Technology),Panneer Selvam Venkatesh Kumar(Toyohashi University of Technology),OKAMOTO Shunya (Toyohashi University of Technology),SHIBATA Takayuki (Toyohashi

著者名(英語): Ammar GHOUS(Toyohashi University of Technology),Rifat Hussain Chowdhury(Toyohashi University of Technology),Venkatesh Kumar Panneer Selvam(Toyohashi University of Technology),Shunya OKAMOTO(Toyohashi University of Technology),Takayuki SHIBATA(Toyohashi University of Technology),Moeto NAGAI(Toyohashi University of Technology)

キーワード: Single-cell Screening|Photocuring|Deep learning|LCD 3D printer|Cell recognition|Single-cell Screening|Photocuring|Deep learning|LCD 3D printer|Cell recognition

要約(日本語): Recent advances have catalyzed a growing interest in automated single-cell screening using image-guided technologies. This study presents a methodology for selective single-cell photocuring with a 3D LCD printer that utilizes deep learning-based cell detection within heterogeneous populations, providing enhanced precision in single-cell screening and analysis. High-resolution images of cells are captured and processed using a custom-trained YOLOv8 model. A 3D model in STL format is generated from the bounding box coordinates after assigning a fixed depth value to each cell. The method uniquely uses an LCD printer to accurately perform photocuring at specific locations, allowing for intricate modifications of individual cells.

要約(英語): Recent advances have catalyzed a growing interest in automated single-cell screening using image-guided technologies. This study presents a methodology for selective single-cell photocuring with a 3D LCD printer that utilizes deep learning-based cell detection within heterogeneous populations, providing enhanced precision in single-cell screening and analysis. High-resolution images of cells are captured and processed using a custom-trained YOLOv8 model. A 3D model in STL format is generated from the bounding box coordinates after assigning a fixed depth value to each cell. The method uniquely uses an LCD printer to accurately perform photocuring at specific locations, allowing for intricate modifications of individual cells.

本誌: 2024年7月4日-2024年7月5日バイオ・マイクロシステム研究会

本誌掲載ページ: 29-33 p

原稿種別: 英語

PDFファイルサイズ: 1,288 Kバイト