Corning Display Technologies Taiwan has recently tied up with two local startups, Memorence AI and TD42, to jointly develop image recognition technologies and solutions.
The collaboration is part of the initiatives of an accelerator jointly set up by Corning and National Taiwan University’s Taihad Entrepreneurship Center TEC in 2019 with an aim to help local startups develop innovative applications optimizing advanced materials and technologies available at Corning.
Memorence AI is an AI technology developer, focusing on smart visual learning and recognition systems to help enterprises improve product quality control and operating efficiency.
DT42 is a specialist for AI technology and embedded systems. It focuses on image recognition solutions on edge devices without the need of connecting with the Internet or uploading related images to the cloud, which saves data transmission costs while keeping data privacy.
The cooperation with Memorence AI and DT42 calls for Corning to provide its backend processing machinery as a testing field for the two AI-centric startups to develop related solutions to assist Corning to improve its image and defective inspection processes, said Daniel Tseng, president of Corning Taiwan.
The collaborative project is divided into two phases. The first phase – to verify whether the methodologies of related solutions offered by the two firms can be applied to Corning’s production lines – got off to a good start and was completed recently at Corning’s plant at Central Taiwan Science Park (CTSP) without being interrupted by the coronavirus pandemic, Tseng noted.
The project is now entering its second phase, aiming to push for implementation of related solutions online, with all tests to be completed by year-end 2020.
Corning’s core capabilities are hardware and materials and the company believes that its manufacturing processes can be more efficient if external software resources are introduced into its operations, Tseng said.
Memorence AI CEO Pai-heng Hsiao revealed that his company’s cooperation with Corning has been focusing on setting the definition of the project, the degree of verification and how to present related results.
Memorence AI previously assisted a leading PCB assembler to establish digitalized manufacturing lines and conduct quality inspection utilizing AI smart systems.
Optimizing digital system and AI-centric solutions, the related efforts have helped reduce the defect failure rate at the PCBA company to below 0.1% from the previous 3-4%, Hsiao revealed.
Meanwhile, DT42 has launched its DLBox (Deep Learning Box), paired with a deep learning edge device software service platform – the DL Edge Platform – enabling edge computing devices to conduct deep learning image recognition. The DLBox is scalable enabling businesses to trial internal customized AI models, according to company co-founder and CEO Jocelyn Chen.
DT42 has also introduced anomaly detection solutions that are suitable for factories with standardized manufacturing processes fulfilling small volume orders in multiple designs, Chen revealed.-Digi Times