Monash University
monash_155886.pdf (4.22 MB)

Characterization, analysis and modelling of vision-based force measurement in microrobotic cell injection

Download (4.22 MB)
posted on 2017-02-27, 02:34 authored by Karimirad, Fatemeh
There has been an increasing demand of micromanipulation in the field of biotechnology during the last decade. One of the important research areas in micromanipulation methodologies and systems is the manipulation of the biological specimens such as the biological cell injection. The limitations associated with the conventional manual cell injection including low success rates, long training curve, risk of contamination and poor reproducibility require for reduction of the direct human involvement in the process. Investigations into the use of visual feedback to establish methodologies for improving the success rate of microinjection process are being carried out. However, concluding from the available outcomes, visual feedback alone may not provide a significant improvement over the existing methods. It has been reported that force sensing capability can directly influence the success rate of microinjection. The direct force information in comparison with the visual feedback can reflect the changes in the physical behaviour of the cell (such as deformation or extent of penetration) more quickly and accurately. The need for force control arises also due to the possible damage to the cell caused by excessive force during the process. However, relatively few research efforts have addressed measurement and control of the interaction forces between the micro-indenter and the cellular surface in the cell micromanipulation processes. This thesis introduces a method to conduct force measurement based on the visual information from the microinjection images extracted by well-established image processing algorithms. This technique is called Vision-Based Force Measurement (VBFM). A methodology is proposed and implemented to provide a robust and real-time strategy to characterize the deformation of biological cells in a quantitative manner during the course of the indentation. The visual data including the deformation, the orientation and the size of the cell obtained from the images serves as the input into an artificial neural network (ANN) model for VBFM. Based on the shape and the geometry of the cell, the ANN model estimates the applied load to the cell as the output. In order to collect the required data to train the neural network, a prototype bio-micromanipulation system with force measurement capability is established. The trained ANN model acts as a precision loadcell to determine the applied load to the cell during microinjection process in real-time. The VBFM neural network (NN) model estimates the force and feeds it back into the system in order to control the position and velocity of the micromanipulator. This NN based VBFM technique alleviates the most of the limitations associated with existing force measurement methods using sensors, which include the size, the integration and the acoustical or thermal sensitivity of the sensors. Furthermore, it doesn’t require additional specified bio-micromanipulation systems, thereby, making it cost effective and flexible. Thus, the proposed methodology can be used to measure forces where it is very difficult or not feasible to use a force sensor such as biological cells microinjection in which cells may be seriously damaged due to the excessive forces. A series of micromanipulation experiments were conducted on zebrafish embryos at various stages of development to verify the effectiveness of the proposed approach. The experimental results demonstrated the capability of the proposed VBFM method with 90.96% repeatability and 97% correlation between the VBFM outputs and the desired outputs.


Campus location


Principal supervisor

Sunita Chauhan

Year of Award


Department, School or Centre

Mechanical and Aerospace Engineering


Doctor of Philosophy

Degree Type



Faculty of Engineering

Usage metrics

    Faculty of Engineering Theses


    No categories selected


    Ref. manager