Aetina to Enable Next-Generation Industrial Inspection Using NVIDIA Metropolis for Factories

NEW TAIPEI CITY, Taiwan, May 29, 2023 /PRNewswire/ — Aetina has launched an AI solution specifically designed to enhance productivity in existing automated optical inspection (AOI) systems in manufacturing. The solution involves NVIDIA Metropolis for Factories and Aetina’s SuperEdge AI platforms, enabling users to reduce the workload of reinspecting the false rejects. Aetina and its parent company Innodisk are now collaborating with NVIDIA to deploy the solution in Innodisk’s factories for advanced electronics manufacturing. 

During Computex 2023, Aetina showcases the AI solution at booth I0810, providing event visitors with a clearer understanding of its functionality.


Aetina to Enable Next-Generation Industrial Inspection Using NVIDIA Metropolis for Factories

Aetina’s AI Solution Based on NVIDIA Metropolis for Factories

The AI solution is based on NVIDIA Metropolis for Factories, which is a comprehensive collection of AI-powered workflows designed to enhance AOI in factory settings. This technology empowers industrial solution providers and manufacturers to develop, deploy, and manage customized quality-control solutions, resulting in cost savings and improved production throughput. The solution incorporates AI software from the NVIDIA AI Enterprise software suite, Siamese network-based AI models, and Aetina’s SuperEdge AI platforms powered by NVIDIA GPUs. It can help prevent factory workers from wasting time on items incorrectly identified as defective in printed circuit board assembly (PCBA) lines using AOI systems for product quality control.

Although AOI is effective in reducing the risk of delivering defective units, there is a chance of misclassifying flawless items as imperfect. This can lead to unnecessary reevaluation during the assembly process. When non-defective items are mistakenly identified as defective, assembly line workers must spend additional time inspecting these items again. In general, there is approximately a 20% chance of units under inspection being identified as defective by AOI. Among the units marked as defective by AOI, some may actually be flawless.

Aetina’s AI solution provides an efficient reinspection process within production line AOI systems. Once implemented, it can identify non-defective items among those that did not pass the initial AOI inspection, significantly reducing the chance of misclassification. After implementation, the chance of units under inspection being marked as defective will be reduced to under 5%, without increasing the risk of misidentifying defectives as defect-free items. This enables assembly line workers to focus their efforts more on fixing genuinely defective items.

Leveraging Innodisk’s AI technology

Aetina developed the NVIDIA Metropolis-based solution by leveraging the AI technologies of Innodisk. Innodisk is a global AIoT solution provider specializing in the seamless integration of hardware and software solutions. The company continues to empower AI transformation across a wide range of industries through its innovative Innodisk AI solution. This transformative solution is delivered through an extensive network of Innodisk’s group companies and esteemed global partners. With the support of Innodisk, Aetina is able to create cutting-edge solutions that bring edge AI to various verticals, including the manufacturing industry.

At the same time, Innodisk is committed to enhancing its factory production processes by adopting next-generation industrial solutions like NVIDIA Metropolis for Factories.

Aetina’s SuperEdge AI Training Platforms

Aetina’s SuperEdge is a series of x86 AI training platforms that are configured with powerful NVIDIA GPUs. Currently, one model from the SuperEdge family has been listed as a member of NVIDIA-Certified Systems 3.0, enabling NVIDIA’s AI software tools to run at maximum performance. Aetina plans to launch more AI training platform models that fully support NVIDIA’s software suite, helping developers quickly create various AI applications.