Reconstruction of missing cells in fluorescent microscopy

Fluorescent microscopy is one of the several types of imaging techniques used by biologists to study cell activities. One challenge of tracking cells from fluorescence microscopy is that cells in fluorescent images frequently disappear and reappear. The situation is further complicated by cell divisions, which also occur frequently in an image sequence. In this paper, we apply a level set method to reconstruct cells that disappear in an image sequence and in particular, cells that are undergoing cell division. The image frames are stacked together to form a 3D image volume. The disappearance of a cell leads to a broken cell path. We reconstruct the incomplete cell paths by a level set segmentation of the 3D image volume. If the disappearance happens during cell division, the level set method segments the visible cell paths before and after cell division, and then joins them together by extending the cell paths into the missing gap. We also propose a simple and cost-efficient method similar to inpainting techniques to capture the cell appearance when it disappears by making use of the level set function obtained from the segmentation. The idea is that the intensities of a visible cell on a level set contour are copied to the corresponding contours of a disappeared cell. We will present results for reconstruction of cells undergoing cell division for C2C12 cells in fluorescent images to illustrate the effectiveness of our method.