A key requirement is properly labelled image data. To better support customers throughout the entire workflow of preparing image data, MVTec Software is expanding the functionalities of its Deep Learning Tool. The tool supports data management, labelling, training, and evaluation of image data. In addition to the continued availability of the free basic edition, two new commercial versions will be released for the first time on 30th April 2025: Deep Learning Tool Professional and Deep Learning Tool OEM.
“As a technology leader in machine vision software, we help our customers take full advantage of the innovative potential of deep learning. With the launch of the new commercial versions of our established Deep Learning Tool, we’re taking the next step toward professionalising the entire deep learning workflow. Machine builders and solution providers can now incorporate the Deep Learning Tool into their solutions and resell it. Additional features and customization options are also available,” explains Christian Eckstein, Product Manager Deep Learning Tool at MVTec.
Labelling: essential for robust deep learning applications
The quality of labelled data plays a key role in the performance, accuracy, and robustness of the application. The Deep Learning Tool already offers a way to label training images in its basic version, which will remain free of charge. Thanks to its intuitive user interface, labelling can be done easily and without requiring programming skills. Prepared data can then be seamlessly integrated into MVTec’s software products HALCON and MERLIC to perform deep-learning-based object detection, classification, semantic and instance segmentation, anomaly detection, and Deep OCR.
The Professional Edition of the Deep Learning Tool, available under a perpetual license, enables customers to incorporate the tool into their own solutions and resell it. With straightforward integration capabilities, machine builders and solution providers can easily add the full power of MVTec’s deep learning workflow to their products without the need of investing in developing their own tools.
The OEM Edition takes it a step further, offering deeper integration, including white-labelling options like custom colour themes, logos, as well as predefined training and evaluation parameters. This makes it possible to tailor the tool to fit a customer’s product or solution, including optimisation for available hardware or specific application requirements.
Both new editions offer customers a cost-effective way to enhance their products with a comprehensive range of deep learning capabilities – including labelling and training. Long-term compatibility with future versions of HALCON and MERLIC is guaranteed, and thanks to regular updates, users remain at the cutting edge of AI technology.
