Deep learning-based software locates rack samples
Clinical diagnostic equipment manufacturers rely on machine vision for automated deck population assessment. Liquid handling systems, such as analyzers or IVD equipment, must detect the location and orientation of racks and their vessels in order to pipette and process samples properly. Incomplete racks can impair workflows and cause costly machine assists. These highly automated systems sometimes encounter failures resulting from misaligned or absent test tubes or caps that haven’t been removed. Vessels and tubes vary in size, dimensions, and cap type by manufacturer, making it difficult for the machine handling system to predict the position of parts on their decks.
Cognex ViDi self-learning algorithms excel at locating similar objects in bulk. With the ViDi Blue-Locate tool, any lab technician is able to fixture tray images, teaching the software to identify and count individual tubes based on their size, shape, and features. ViDi Blue-Locate detects the location and orientation of microtiterplates and other racks on the IVD’s deck, as well as their sample tubes, despite their unpredictable and varying dimensions. Once ViDi has located the samples on the rack, traditional machine vision tools are used to align the samples and reagents for further processing.