Synaptrix Parts Inspection (SPI) is a Michigan-based LLC formed to exploit IEI’s revolutionary machine vision applications within both industry and government. The company specializes in performing quality control, as well as parts classification using IEI’s patented neural network paradigms that include the Self-Training Artificial Neural Network Object (STANNO) and the Creativity Machine Paradigm.

What Makes Synaptrix Vision Applications so Unique?

Until recently, vision systems that perform parts inspection and/or classification require much tedium and time to build. Oftentimes, technicians must identify key part features and their geometrical interrelationships. These explicit rules are then embedded within traditional computer code. Oftentimes, this painstaking process requires days or even weeks to perfect, and even then, such hard-earned part recognition filters are susceptible to variations in lighting conditions and background content. Furthermore, these computer codes only recognize the part when presented at some favored aspect angle.

In remarkable contrast, IEI has developed and patented a neural network scheme that in seconds automatically forms recognition and anomaly filters that are robust to lighting, background, and aspect angle variations. These systems are trained by simply moving the object, through a wide range of orientations in front of a camera linked to IEI neural networks that are training in real time. Parts and/or defects may be rapidly classified using these totally autonomous neural architectures, over broad dynamic ranges of illumination, background types, and viewing perspectives.

What Kinds of Projects is SPI Involved In?

SPI has responded to a number of heretofore impossible challenges presented by the Detroit automotive industry. Notable among these projects have been the detection of anomalous placement of engine parts as well as consistency checks between various stabilizer bars and their attached ID codes. SPI is also in the process of prototyping a plumbing parts classification scheme to visually identify and separate, intermingled types of fittings arriving via conveyor belt.

Freely moving plumbing fittings, such as this adaptor, are automatically classified after just a few seconds of neural network training…

To learn more about Synaptrix Parts Inspection, contact: