Image Processing and Analysis for Applications in Biological Systems We are currently working on an EPSRC funded collaborative project in this area. The objective of the project is to research and develop a novel particle tracking system for live cell fluorescence images. The present work involves both image denoising and particle tracking. Our collaborator is Ilan Davis Research Group at the University of Oxford. Image Denoising We have developed a new denoising approach that combines feature extraction and non-local means filtering for the analysis of live cell fluorescence images. The algorithm works by feeding class information, extracted using the Haar-like features, to an adaptive non-local means filter so that the latter can adapt to the changes between noisy background and particles. As such, the algorithm has additional capacity to balance background smoothing with particle enhancement, and requires no prior information about noise distribution within the image. Experiments show that the algorithm not only achieves considerably higher PSNR than several commonly used denoising filters, but also performs much more efficiently in background smoothing of live cell fluorescence images when compared to non-local means filtering and patch-based denoising. Particle Tracking The quantitative analysis of particle trajectories provides important information about working mechanisms and structures in living cells. Conventional particle tracking methods are based on nearest neighbour association or Bayesian state estimation. But for low SNR live cell fluorescent images, these two ideas alone commonly can lead to problems. We are working on combining these two methods for robust particle tracking in live cell imaging sequences. Figure: Particle tracking: a sample To be continued …