Easy labeling of image data

The machine vision industry is gaining importance through the use of deep learning, a form of artificial intelligence, as it enables the automation of entirely new applications and also improves results. A prerequisite for this is cleanly labeled image data. To better support customers throughout the entire image data preparation workflow, MVTec Software GmbH is expanding the functionality of its Deep Learning Tool software product. It enables data management, labeling, training, and evaluation of image data. In addition to the basic edition, which will continue to be free, two commercial versions will be available for the first time: Deep Learning Tool Professional and Deep Learning Tool OEM.

"As one of the technology leaders for machine vision software, we want to help our customers make the most of the diverse advantages and innovative power of deep learning. With this in mind, we are taking the next step toward more professionalization of the entire deep learning workflow with the new, commercial versions of our proven deep learning tool. Machine manufacturers and solution providers of machine vision applications now have the opportunity to make the deep learning tool part of their solution and resell it. In addition, additional functions and customization options are available," Christian Eckstein, Product Manager Deep Learning Tool at MVTec, summarizes the new features of the deep learning tool.

Labeling is essential for robust deep learning applications

Effective labeling of training data is an essential step in any deep learning application. The quality of the labeled data plays a crucial role when it comes to the performance, accuracy, and robustness of the application. The deep learning tool already offers a way to label training images in the basic version, which will remain free of charge. Thanks to the intuitive user interface, this step is easy and requires no programming knowledge. The prepared data can then be seamlessly integrated into the two MVTec software products Halcon and Merlic to perform deep learning-based object detection, classification, semantic and instance segmentation, anomaly detection, and deep OCR.

The paid Professional Edition of the deep learning tool allows customers with a non-expiring license to make the tool part of their own solution and allows for resale. With these simple integration options, machine builders and solution providers can easily add the full power of the MVTec deep learning workflow to their product without having to invest in their own tools.

The OEM Edition goes one step further and allows for even deeper integration, for example, by enabling white-labeling options such as color theme configuration, custom logos, and presetting training and evaluation parameters. This edition can be individually tailored to the customer's product or solution by adjusting the parameters to the available hardware or use cases.

With both new editions, customers benefit from a manageable investment, allowing them to quickly upgrade their product with a full range of deep learning functionalities, including labeling and training. Long-term compatibility with future Halcon and Merlic versions is naturally guaranteed. And thanks to continuous updates, users are always at the cutting edge of technology in the fast-moving world of AI.