SEOUL, March 31 (Yonhap) -- Hyundai Mobis Co., South Korea's biggest auto parts maker, said Sunday that it aims to secure an advanced deep learning-based camera image detection technology by the end of this year.
The company said it plans to apply the technology, required to build safe self-driving systems, to the front camera sensor of an autonomous vehicle from 2022 for mass production.
"We are targeting deep learning-based information and communication technology to differentiate our autonomous driving and connectivity technologies," Jang Jae-ho, head of the Electrical & Electronics R&D Center of Hyundai Mobis, said.
Deep learning-based object detection technology can accurately recognize not only distant objects but also overlapping ones. It can distinguish one object from another -- for instance, between vehicles and pedestrians or between pedestrians and bicycles -- and utilizes the information as data for safe driving, the statement said.
It can, moreover, read the movement patterns of vehicles and the posture of pedestrians to predict their actions.
To evolve into a software-centered technology company, Hyundai Mobis has made a concerted effort to invest in competitive companies in the fields of futures vehicles.
It poured 8 billion won (US$7 million) into a stake in local startup StradVision, which possesses deep learning-based camera image detection technology, and invested 5.5 billion won into a stake in Chinese computer vision startup Deep Glint.
Seoul-Tokyo ties tipped for deeper rift after Japan's expanded export control: experts
Trade row with Japan, another headwind for Korean economy
Japan hides true intentions, leaving export curbs unjustified
Boycott of Japanese goods to intensify as Tokyo expands export curbs
Japan's broader export curbs aim for S. Korea's Achilles heel