Confocal microscopy is used widely for 3D biological imaging, but can be too slow for many applications. The limitations arise from scanning a single spot across the specimen at high speeds. Singe-spot confocal imaging usually works at a 1-2 Hz frame rate and faster systems tend to be signal/noise limited. Many live cellular events require both high speed and high SNR. Some parallel confocal systems have been developed to collect light from many points simultaneously to obtain high SNR and/or speed. Nipkow disks are compared of many pinholes, but have a fixed pattern and low light efficiency. Slit scanning systems collect an entire video line at a time, but compromise resolution. The TI Digital Micromirror Device (offers an alternative via large arrays of rapidly re-configurable micromirrors that can form arrays of reflection 'pinholes'. The prototype presented here exhibited 0.4 X 0.4 X 0.8 micrometers 3 resolution with a 100X 0.90 NA objective. An array of 10,000 or more neighborhoods, each compromising a single ON mirror in a group of OFF mirrors, creates the confocal parallelism. Alternating the ON mirror in each neighborhood until the image is completely formed on the CCD sensor enables transverse scanning. With 10,000-fold parallelism, for example, light collection efficiency and frame rate can both be 100X higher than in typical spot scanning. The high sensitivity allows high-speed confocal imaging at intensities below the cellular fluorotoxicity threshold. This was demonstrated in a hamster window preparation scanned daily in one-week longitudinal studies. Vessel geometry and localized blood flow were reconstructed to measure perfusion. High frame rate and sensitivity allowed real-time visualization of DiI stained intravascular red blood cells with no apparent tissue damage, supporting the tremendous potential advantages over current confocal technologies.