Retracted: Localizing and extracting filament distributions from microscopy images

摘要: Summary Detailed quantitative measurements of biological filament networks represent a crucial step in understanding architecture and structure cells tissues, which turn explain important events such as wound healing cancer metastases. Confocal microscope images specimens marked for different structural proteins constitute an source observing measuring meaningful parameters networks. Unfortunately, current efforts at estimation orientation from microscopy are predominantly limited to visual indirect experimental inference. Here we describe new method localizing extracting distributions 2D confocal images. The combines filter-based detection pixels likely contain with constrained reverse diffusion-based approach the filaments centrelines. We show qualitative experiments, using both simulated real data, that can provide more accurate centreline estimates comparison other approaches currently available. In addition, algorithm is robust respect variations initial often used. demonstrate application experiment seeks quantify effects carbon nanotubes on actin cytoskeleton live HeLa cells. their presence disrupt overall cytoskeletal organization