Version 17.12 of MVTec’s HALCON software features an extensive set of deep learning functions, as a new set of self-learning algorithms reportedly enables users to simplify and accelerate programming processes significantly while being able to benefit from more robust detection rates and better classification results. This includes a range of functions for training convolutional neural networks (CNN).
MVTec also brought its deep learning functionalities to embedded boards with NVIDIA Pascal architecture in version 17.12, which ships with two pretrained networks, one of which is a "compact" network that is optimized for speed and suitable for use on embedded boards. This software was successfully tested on NVIDIA Jetson TX2 boards based on 64-bit ARM processors. The deep learning inference, i.e., applying the trained convolutional neural network 9CNN) almost reached the speed of a conventional laptop GPU (approx. 5 milliseconds). This, according to MVTec, is an unusually-high performance for an embedded device compared to a standard PC.
Now, users can deploy deep learning techniques on the NVIDIA Jetson TX2 embedded board. MVTec will provide interested customers with a software version for this architecture on request.
The Vision Show 2018 booth number: 302
To Learn More:
Headquarters: Munich, Germany
Product: HALCON and MERLIC machine vision software
Key Features: Deep learning functions (including a range of functions for training convolutional neural networks), pretrained network optimized for use on embedded boards.
Share new products that you think are particularly interesting or helpful by contacting James Carroll, Senior Web Editor, Vision Systems Design.