Curvilinear networks in cells and tissues are routinely imaged by fluorescence microscopy and sometimes optical coherence microscopy (OCM). Gaining insight into the structural and mechanical properties of these curvilinear networks and understanding the mechanisms of their formation require image analysis methods for automated quantification of massive image datasets. The diversity of these networks as well as low image quality make reliable extraction challenging. This chapter presents a recent body of work that provides a complete framework for extracting the geometry and topology of multidimensional curvilinear networks in microscopy images. The proposed multiple Stretching Open Active Contours (SOACs) are automatically initialized and evolve along the network centerlines synergically: they can merge with others, stop upon collision, and reconfigure with others to allow smooth continuation across network junctions. The approach is generally applicable to 2D/3D images of curvilinear networks with varying signal-to-noise ratio (SNR). Qualitative and quantitative evaluation using simulated and experimental images demonstrate its effectiveness and potential.
|Original language||English (US)|
|Title of host publication||Biomedical Image Segmentation|
|Subtitle of host publication||Advances and Trends|
|Number of pages||30|
|State||Published - Nov 17 2016|
All Science Journal Classification (ASJC) codes
- Biochemistry, Genetics and Molecular Biology(all)