A major highlight in the new HALCON 21.05 version is the Generic Shape Matching feature, which makes MVTec's industry-proven shape matching technologies even more user-friendly and future proof. Users can now implement their solutions much faster and more easily, thanks to a significant reduction in the number of operators required. Moreover, thanks to the unification of HALCON's different shape matching methods into a single set of operators, users can now integrate new shape-matching-related features more easily. The HALCON Deep Learning Framework is another new feature that enables experienced users to create their own model within the software. As a result, experts can now implement even the most demanding and complex applications in HALCON directly, without being dependent on pretrained networks or frameworks from third-party suppliers. HALCON 21.05 also supports the NVIDIA® TensorRT™ inference engine, which can significantly improve the inference performance of applications. In addition to the speedup, users also benefit from significantly lower memory consumption, giving them greater flexibility in their choice of hardware.