Acquisition of 4D in vivo confocal data from anesthetized mice is a complex task that involves overcoming a myriad of challenges such as proper surgical preparation, monitoring of vital physiological parameters of the anesthetized animal and careful selection of correct fluorescent antibodies and acquisition parameters. Also, it often requires mechanical stabilization of the tissue being imaged. Despite stabilization of the target tissue, the breathing of the mouse can sometimes induce artifacts of the image currently being acquired. In confocal image stacks, each breath manifests as a distorted image plane in the stack, and commonly every stack has several distorted planes (white arrows).
The presence of artifacts interferes with the registration of stacks as well as segmentation and tracking of cells over time. Our goal is to together with BIIF develop a pipeline for automatic detection of distorted image planes in 3D-timelapse datasets and the subsequent replacement of the affected planes by interpolation of neighbouring planes in the stack.