Deep learning used to detect problematic products
South Korean mobile carrier LG Uplus announced this week that its artificial intelligence technologies would pinpoint defective products without human intervention.
The Seoul-based company noted that it had built a smart factory for Dongjin Techwin, which churns out electrical equipment and appliance in Changwon.
LG Uplus said that its “AI vision inspection” system would perform the final quality inspection tasks of dryer cabinet covers. Thus far, the works have been carried out manually.
The new system would be powered by deep learning, a machine learning training model developed in imitation of the decision-making fashions of the human brain.
After learning a set of structured image data, the new system is to pinpoint faulty products only based on visual AI inspection.
Dongjin Techwin hopes that the new inspection system will be able to find defects, which human eyes failed to detect.
“The AI vision inspection is expected to find out inferior products and provide data necessary to advance the automation procedures,” Dongjin Techwin official Jeon Geun-sik said.
LG Uplus launched the smart factory business last year for small- and medium-sized enterprises.
Experts point out that deep learning brings innovations to brick-and-mortar businesses.
“Manufacturers have been slow in adopting AI technologies,” Prof. Seo Yong-gu at Sookmyung Women’s University said.
“Yet, the development of AI technologies makes even traditional manufacturing corporations adopt them. Such a trend is expected to speed up.”