LG Innotek announced on June 10th that they have successfully lowered the defect rate of high-spec camera modules by using AI technology. With the upcoming use of AI technology, LG Innotek is solidifying its position as the top global manufacturer of camera modules by streamlining the production of intricate components and reducing the need for lengthy Ramp-up processes. LG Innotek aims to enhance profitability through the implementation of advanced 'Industrial AI' technology, as well as improving cost competitiveness by stabilizing the initial yield.

Low production output in the initial phases of large-scale manufacturing of valuable new products is a frequent problem in the manufacturing industry and poses a considerable obstacle. Manufacturers repeat simulations and test productions numerous times until they find the right process recipe capable of producing high-quality products. The longer this process takes, the higher the F-Cost (cost incurred to correct a defect after mass production) and the more delayed mass production becomes.

Failing to meet the delivery date results in a decrease in customer trust. This is why component manufacturers place so much importance on stabilizing the initial yield, sometimes even risking their business. LG Innotek believes that the key to tackling this challenge is through AI, leading them to start developing AI that can forecast process defects starting in 2021.

AI examines the entire initial process and identifies where defects are likely to occur before they happen. AI uses defect prediction results to create a simulation that considers various factors in the process, such as the strength of the machinery, the speed of the conveyor belt, and the indoor temperature. By doing this, a process recipe that can reduce the defect rate to a minimum is developed.

Last year, LG Innotek used an 'AI process recipe' for the first time in the industry to manufacture a high-spec camera module. Consumers prefer smartphones that are slim, lightweight, and offer enhanced performance. The thickness and size of the high-spec camera module are expected to be reduced in line with this trend.

The manufacturing process becomes increasingly sophisticated as the camera module specifications become ultra-slim with high specifications. The camera module undergoes numerous stages of highly precise processes. Previously, when a flaw was found during performance testing at the start of camera module mass production, it would take over 72 hours to create a new process recipe.

Nevertheless, by implementing the 'AI process recipe', the process has been condensed to just 6 hours. This is no longer an issue due to the elimination of the need to repeatedly make recipe corrections and sample productions based on an engineer's experience. The defect rate in major inspection items, which previously had a high defect rate, has been reduced by up to 90% due to the implementation of the 'AI process recipe.'

According to a spokesperson from LG Innotek, they utilized AI to analyze tens of millions of data points on the camera module process in order to develop the best possible manufacturing method. Additionally, they mentioned that these data resources will be crucial in enhancing the company's quality standards. Effective in enhancing cost competitiveness--AI will be applied to semiconductor substrate process--AI will be applied toonductor substrate process.

LG Innotek was able to stabilize the initial yield of camera modules and significantly enhance cost competitiveness by utilizing 'AI process recipe'.