CellSnake: A Vision-based Methodology to Dynamically Track Cell Deformation during Cell Micromanipulation
There has been an increasing demand in the field of micromanipulation during the last decade. One of the important applications of micromanipulation methodologies and systems is the manipulation of biological specimens. In this paper, the focus of attention is on the cell injection application. A biological cell is very delicate and is deformed as soon as touched by the head of a needle. In a manual operation, hand tremor may seriously damage the cell during manipulation, thus an automated system is required to perform cell injection. Successful autonomous and/or automated cell injection requires robust algorithms for image segmentation and visual cell tracking. The speed of visionbased algorithm strongly depends on the speed of localization and recognition procedures. The proposed vision-based methodology provides a self-contained and real-time strategy to recognize a cell, and track its deformation during injection. An algorithm is established to acquire the boundary of the cell. In order to improve process speed and accuracy, it is required to remove other objects such as pipettes and the median impurities from the scene of injection. Therefore, contour extraction, object recognition, and localization are the issues considered in the proposed system. The methodology also supports real-time process by applying the above algorithms successfully at realtime frame rates.
Fatemeh Karimirad Bijan Shirinzadeh Umesh Bhagat Yongmin Zhong Sergej Fatikow Julian Smith
Robotics and Mechatronics Research Laboratory (RMRL), Department of Mechanical and Aerospace Enginee Department of Mechanical Engineering, Curtin University of Technology, Perth, Australia Department of Computing Science, University of Oldenburg, Oldenburg, Germany Department of Surgery, Monash Medical Center, Faculty of Medicine Nursing and Health Sciences, Monas
国际会议
长春
英文
1-5
2011-08-29（万方平台首次上网日期，不代表论文的发表时间）