Advances in image-based quantitative measurements in biological imaging

The greatest innovations in microscopy now occur after image acquisition. Digital image processing techniques continue to improve microscopy driving scientific discovery in the 21st century. This thesis presents original knowledge in the development of advanced digital processing methods for quantification of the three distinct physical attributes of shape, motion and force. A key area in microscopy is imaging of biological samples with low contrast. This thesis presents an alternative three-dimensional tomographic microscopy method, termed Refractive Index Tomography (RIT), which uses the refractive properties of the specimen for imaging without the need for fluorescent labels. RIT uses a straightforward, low-cost optical setup with an algorithm applicable to both visible light and X-ray-based microscopy, further demonstrating the versatility of software over hardware in the modern microscope. Motion and flow are increasingly being understood to play a large role across a wide range of biological processes. Particle Image Velocimetry (PIV) is an image-based technique developed for fluid dynamics research, and has recently gained interest amongst biological researchers for studies involving the quantification of biological motion and flow. However, limitations in optical setup and access to the region of interest compromises image quality through low light, low contrast and uneven illumination. This thesis presents two processes for refining PIV by addressing these issues. The first is a hybrid averaging process that combines and optimises two averaging methods to reduce the effect of image noise on flow measures. The second is Polynomial Element Velocimetry, a completely new variant of PIV designed for quantifying flow and flow gradients in complex biological flows from images with limited resolution and contrast. Quantification and characterisation of the role of biomechanical forces in early embryonic development is of increasing interest to researchers. This thesis develops a novel technique to quantify cortical tensile forces in living mouse embryos. Four-dimensional cell segmentation and tracking were used to demonstrate that inter-cellular forces within the embryo are required for the formation of viable embryos. This critical advancement in evidence suggests mechanical forces may be used to control pluripotency of embryonic stem cells needed for tissue regeneration and treatment of disease. Advances from image processing software have enabled the capture of dramatically more functional information over what is possible with hardware alone. The studies presented in this thesis represent developments in image processing that will continue to drive digital microscopy, from visualisation of structures into an increasingly powerful tool for quantification of physical processes.