Deep learning is an AI technology that systematizes the recognition of images, videos, sounds, and languages in a way that could previously only be achieved using human sensory organs. Highly precise, it is discriminating in a way other computers are not and possesses a human-like flexibility. Using this technology, inspection targets are processed while learning occurs, and the more learning occurs, the more accurate the system becomes. Our 2D (image) inspection system, Deep Inspection, and 3D (object/space) point cloud data recognition system are mainly applied to manufacturing and medical fields.
Take, for example, quality checks performed in the manufacturing industry. Conventional image inspection systems cannot be said to be highly accurate, meaning that human beings have to conduct the final visual inspections throughout the day. However, we believe the human eye is better used looking up at the sky rather than straining to find small scratches. While it is a simple, natural thing, the fact remains that there are people who damage their eyes through visual inspection. AI helps to relieve that burden.
The reality is there is no existing AI capable of performing “perfectly.” However, our goal is to minimize the burden on working people as soon as possible. It is for that reason that Rist applies deep learning. Deep leaning enhances AI capabilities through multi-layered processing and uses our original algorithm to assign a numerical confidence score. Items with lower scores are held back, and the final judgment is left to a person.
We also developed a comparative inspection solution that can learn permissible differences from a single good image, and prevent over-detection. In this way, AI performs the bulk of the inspections, and humans visually inspect only those low confidence score items that are put on hold. Thus, it is with the AI + person tag that the burden on the human worker is reduced while limitlessly increasing inspection accuracy.
With our core belief in “making people happy with the power of science,”
Rist will continue creating AI to provide more freedom to people in a variety of fields.
