IDS - Using multiple neural networks in one image
The embedded vision system 'IDS NXT' has new features. These include multi-ROI (Region Of Interest) for AI-based object detection and the option to use different neural networks for different ROIs in one image via the vision app, as well as binning, linescan mode and performance and configuration improvements. New customers can also choose whether they want to use the IDS NXT lighthouse training software on AWS as previously offered by IDS or via Microsoft Azure Cloud. With Microsoft Azure Cloud Services, IDS Imaging Development Systems now supports another host for the AI training software IDS NXT lighthouse in addition to AWS. The offer is aimed in particular at companies that are not allowed to work with AWS due to regulatory requirements. The range of functions and services as well as the operation of the training software is the same for both hosters. Instead of setting up their own development environment, users can start training their own neural network directly in IDS NXT lighthouse - even without prior knowledge of deep learning or camera programming. This involves three main steps: Upload sample images, label the images and then start the fully automatic training. The generated mesh is optimized for use with IDS NXT industrial cameras.
IDS NXT is a complete AI-based vision system consisting of intelligent cameras with a software environment that covers the entire workflow from the creation to the execution of AI vision applications. The software tools make AI-based image processing usable for different target groups without prior knowledge of artificial intelligence or application programming. In addition, professional tools enable open-platform programming, which means that the intelligent IDS NXT cameras can be designed and used very individually.