In the era of Industry 4.0, manufacturers face ever-increasing pressure to improve quality control while reducing operational costs and minimizing human error. Traditional inspection systems often fall short when handling complex defect detection tasks or adapting to production line variations. To meet these evolving needs, deep learning-based vision inspection powered by scalable edge AI platforms has emerged as a transformative solution. With advancements in hardware performance and no-code AI software, intelligent inspection systems can now be seamlessly implemented on compact edge devices—delivering real-time accuracy while reducing reliance on cloud infrastructure.
Portwell, a global industrial embedded computing provider, joins forces with Neurocle, an AI vision software innovator based in South Korea, to deliver high-performance edge AI visual inspection solutions tailored for manufacturing. At COMPUTEX 2025, this partnership showcased a powerful combination of Portwell’s NVIDIA® Jetson modules-based PJAI series hardware and Neurocle’s intuitive no-code deep learning vision software suite. The joint solution empowers manufacturers with scalable, operator-friendly inspection capabilities that effectively address the challenges of traditional systems. By integrating Neurocle’s automated deep learning models onto Portwell’s edge compute platforms, users can achieve up to 2.5x faster inference with real-time defect detection—without needing dedicated AI coding engineers.
Neurocle brings an advanced AI software stack built for industrial visual inspection. Its flagship product, Neuro-T deep learning model trainer, offers a no-code, GUI-based interface that allows users to build high-performance deep learning models through automatic architecture and hyperparameter optimization. Complementing this, the Neuro-R Runtime Library enables seamless integration of trained models into production lines, supporting real-time inspection across diverse industrial environments. Neurocle supports nine vision model types—including classification, segmentation, object detection, and anomaly detection—making it versatile for a wide range of applications from semiconductor to battery production. With 99.9% inspection accuracy and 1.7ms inference speed, Neurocle sets a new standard for usability and performance.
Portwell’s edge computing solutions are designed for industrial-grade AI deployment. The compact PJAI-100 series, equipped with the NVIDIA Jetson Orin™ NX module or Jetson Orin Nano™ module, delivers up to 100 TOPS AI performance in a fanless system build—ideal for factory automation and computer vision use scenarios at the edge. Moreover, the PJAI-200-AO32, powered by the NVIDIA Jetson AGX Orin™ module, boasts up to 200 TOPS and supports high-bandwidth I/O interfaces with 12 PoE ports and 10GbE connectivity—perfect for high-throughput inspection tasks like pouch-type battery production. These platforms ensure thermal reliability, robust operation, seamless integration, and agile scalability for edge AI workloads in space-constrained environments.
Together, Portwell and Neurocle offer a powerful AI inspection framework optimized for real-time intelligence in real-world deployments. In automotive applications, the solution excels at part assembly validation and X-ray tire defect detection—enhancing safety and production efficiency. For the battery sector, it supports both surface inspection and internal X-ray analysis of cells—crucial for ensuring performance consistency and preventing critical failures. With Neurocle’s no-code AI software and Portwell’s edge AI embedded computing platforms built with the NVIDIA Jetson modules, manufacturers can rapidly deploy and scale intelligent inspection systems—improving yield rates and minimizing operational overhead.
The Portwell-Neurocle partnership exemplifies the future of industrial automation: intelligent, flexible, and accessible. By combining best-in-class edge AI hardware with automated no-code deep learning software, the joint solution lowers barriers to AI adoption while delivering measurable performance gains. At COMPUTEX 2025, this collaboration demonstrated a shared commitment to “AI-Next” innovation, empowering manufacturers to embrace smart factory transformation and accelerate their journey toward autonomous quality control to ultimately elevate business outcomes.