Accelerated Inference for Deep Learning in HALCON

With the AI² interface (AI Accelerator Interface), MVTec provides a high-performance and flexible solution for running deep learning models on specialized hardware. This significantly accelerates inference (executing trained neural networks), including on edge devices with limited compute resources.

The AI² interface extends HALCON with a unified programming interface that enables the use of different hardware platforms. This allows users to distribute compute load effectively and use hardware resources efficiently – from high-performance GPU systems to energy-efficient embedded systems.