HOME / CYTH PRODUCTS / NEURALVISION
NeuralVision is a machine vision platform at the forefront of Deep Learning and Artificial Intelligence like abilities applied to the industrial inspection space.
For the first time companies are able to have total control of the performance of their machine vision systems and not be dependent on external vision experts to make changes or incorporate new product lines.
Traditional machine vision is highly dependent on having a controlled environment, rigid positional tolerances, and ultimately the skill of the vision programmer.
It is up to engineers to come up with every algorithm required to inspect a part from measurements to color to correct locations and everything in between.
But what if they miss something?
What if there are parameters they don’t know matter?
This has always been the struggle in the traditional space which results in massive amounts of time going into testing and troubleshooting a huge array of possibilities.
Engineers need total control of part presentation, lighting, repeatability, and to simply always know what’s coming.
NeuralVision from Cyth Systems was designed to allow a person with no machine vision experience to inspect and classify products. Machine vision systems traditionally work by having an experienced programmer choose one of many analysis algorithms to apply to an image.
A programmer might use hole detection, temperature analysis, or width measurement to inspect an image and determine good or bad parts.
NeuralVision applies these same operations in millions of ways behind the scenes looking for any recurring pattern or key measurements that identify a good or a bad product.
The system can recognize the shape and temperature of cooked bacon, or the proper height and shape of croissants, even in cases where human eyes or programming may not be able to define a product’s key parameters.
"Cyth is a critical supplier for us. They're involved in the design, building, and supporting automation tools through our manufacturing."
-J.N., Semiconductor Equipment Manufacturer
It works the same way we learn to recognize an apple from an orange. As people see objects for the first time they unconsciously are looking for unique features, similarities, and differences between all the different objects observed.
It isn’t necessary for you to know exactly what all those key features are in determining an apple, orange, or any other object. What you care about is that you are able to identify the various items as well as notice possible issues within each object.
NeuralVision works in much the same way. As the system is fed images of the relevant objects and told either what unique parts are or that what it is seeing is good or bad it will apply millions of algorithms to learn to identify what is being seen.
By showing the system a wide range of variation such as lighting, shadows, and environments the system will learn to understand what features are important or not important to identifying a part.
An example of this is if you see an apple that is new, old, red, or green you still know that it is an apple. If NeuralVision were shown all these different scenarios and told they were all an apple it would learn to understand all these different examples and what makes an apple unique.
NeuralVision puts the power in the company’s hands. Anyone with a knowledge of their product could look at collected images and begin to identify what parts are or what is good and bad.
This knowledge is then processed to teach the system to think as the operator thinks and replicate their expertise. Here is where the strength of NeuralVision comes in. If a system can be taught by the best production line experts to be as reliable as they are and also never get tired, you’ve just hired the best employee ever!
What's next in
Reach out to Cyth to go over your needs and specific products. We will design a vision system unique to your requirements and facilities to collect images and eventually work as your automated solution.
Cyth will teach operators on the use of the system to collect and then begin identifying images. After this initial collection is done Cyth will handle the Deep Learning training of the system. As the system goes through a few cycles of learning and testing you’ll be ready to go for production. Now you have the knowledge and abilities to train more or new parts as needed!