Always at the forefront of innovation, Data-Pixel has designed a new microscopy method for measuring defects and scratches using deep learning algorithms. This is a major advance in the industrial vision applied to fiber optic connectors which allows diagnostic performance superior to standard software. A patent has been filed to protect this innovation.
Artificial intelligence at the service of diagnostics
Deep Learning is a family of software that belongs to Machine Learning, itself belonging to artificial intelligence. Deep Learning software learns from a set of representative samples and then diagnoses your product using taught criteria. This continuous learning process is carried out by Data-Pixel using images of characterized defects and scratches.
Precise and scalable software
Excellent learning technology, the software is trained by Data-Pixel from a base of defects and scratches and continually evolves. If you find that the analysis is not relevant to you process requirements, it is possible to continue teaching the software from provided samples. Data-Pixel will then optimize the abillity of the software so that it becomes more and more efficient and relevant for future diagnostics.
BLINK WITH DEEP LEARNING
Data-Pixel Deep Learning solution is integrated into BLINK thanks to a simple software update. Thanks to this new functionality, when passing your mouse over a scratch or a defect, Blink will provide you with detailed information such as the types, areas, sizes and scores of these defects and scratches. The Deep Learning establishes a scoring of the defects and scratches based on a learned data set enabling a fine adjustment of the detection threshold sensitivity to meet your inspection level requirements.