Deep learning functions on embedded devices

MVTec, the software provider for industrial image processing, focuses in particular on speed, robustness and hardware compatibility for use in all industrial sectors. Visitors can experience the features of the current releases of the standard software products MVTec Halcon and Merlic live. The manufacturer will also provide information about the advantages of professional, commercial machine vision software and give an insight into the extensive application possibilities of deep learning technologies on embedded devices.

Live demos illustrate practical application scenarios

A number of live demos offer a practical insight into embedded vision applications: a multi-platform setup, for example, shows how four embedded boards solve various tasks live with Halcon and Merlic. This clearly shows the variety of platforms on which the MVTec software runs. Another demo illustrates how the Merlic software identifies and checks medical sample tubes as objects. The Pallas smart camera from MVTec's Chinese partner Daheng Imaging is also used here. Finally, in a third demonstration, five standard deep learning examples will be executed in parallel, namely semantic segmentation, object detection, classification, optical character recognition (OCR) and MVTec's latest feature, anomaly detection. An Nvidia Xavier will serve as the platform for this.

The software provider will also be taking part in the trade fair's supporting program: Christoph Wagner, Product Manager Embedded Vision, will be giving a presentation on "Why Choosing the Right Machine Vision Software for an Embedded Vision Product is not a No-brainer" at the embedded world Conference on February 27 from 10:30 to 11:00 am. MVTec will also participate in the accompanying Embedded World exhibitor forum on February 25 from 1:30 to 2:00 p.m. with a presentation on "MVTec Halcon Deep Learning for Embedded Devices".

Embedded World: Hall 4, Booth 203