Collaboration Boosts Deep Learning Speed for Embedded Vision Systems
MVTec strengthens its machine vision software by integrating high-performance neural processing hardware, enabling faster and more efficient deep learning execution on embedded smart cameras.
www.mvtec.com
How MVTec’s Collaboration Enhances Deep Learning Performance on Embedded Vision Devices
Companies deploying machine vision increasingly require faster and more power-efficient deep learning execution on compact, embedded platforms. To meet this demand, MVTec expanded its technology partner ecosystem through a collaboration with Qualcomm Technologies, enabling its HALCON machine vision software to take advantage of the advanced neural processing capabilities of the Dragonwing processor family.
This integration provides HALCON users with access to a high-performance neural processing unit designed for industrial image processing. It allows smart cameras and embedded vision systems to process complex deep learning tasks at significantly higher speeds while maintaining low power consumption—two essential requirements for modern industrial automation environments.
Interface Development for High-Efficiency Machine Vision
As part of this collaboration, MVTec developed an interface connecting its HALCON software with the Dragonwing RB3 Gen 2 platform running on Qualcomm’s upstream Linux and AI framework. This interface enables accelerated inferencing performance for applications such as defect detection, object classification, and feature localization, which typically require substantial computational resources.
The processor platform supports up to 12 dense TOPS, allowing vision applications to operate efficiently within size- and power-constrained smart camera designs. This performance improvement directly benefits manufacturers aiming to scale Industry 4.0 automation without depending on large PC-based systems.
Supporting the Evolution of Deep Learning in Machine Vision
Deep learning has become a core element of modern machine vision, enabling higher accuracy and automation across industrial sectors. MVTec’s Embedded Vision initiative ensures its software runs optimally on a wide range of compact hardware architectures. By integrating high-performance neural computation through this partnership, MVTec enables customers to improve existing machine vision workflows and develop new automated processes more efficiently.
The company’s Technology Partner Program promotes interoperability within the image-processing ecosystem, bringing together innovative component providers to ensure consistent performance across hardware platforms. The collaboration with Qualcomm Technologies aligns with this objective by combining advanced embedded AI hardware with MVTec’s established vision software portfolio.
www.mvtec.com
