Train segmentation models without programming or AI expertise. Upload your labelled training datasets* into the interface and let VGTRAINER take it from there. You can have your first working model in a single day and can iterate as many times as you like with additional data to improve the model’s accuracy.
*Curate accurate, consistent training data in VGSTUDIO MAX.
Upload your labelled training datasets into the interface and let the software handle the rest. You can have your first working model in a single day and iterate as many times as you like with additional data to improve its accuracy.
“The results are very impressive. After training a model with just 20 labeled datasets, the resulting segmentation is about 98% accurate. It's very impressive, very efficient.”
We know the idea of generating your deep-learning models can feel overwhelming. That's exactly why we developed VGTRAINER: To put the power of AI for segmentation into the hands of everyday users. Check out this video tutorial and see just how intuitive the application is in practice!
Watch the tutorial
Get your free, 4-week trial license and see why we're the industry’s first and longest-trusted CT analysis software.
Request free trial
Put your models to work, again and again! Fine-tuning enables you to adjust an existing deep segmentation model to new but similar inspection tasks with minimal data input in less time.
While you might have needed lots of labeled CT scans to properly train your pore detection model from scratch, you may only need two or three labeled CT scans to fine-tune a properly pretrained model to your specific
task.
Fine-tuning models is ideal for scenarios where you have several production lines to which you need to adapt, or in response to signs of wear and tear. This ensures your models are always aligned with your specific operational needs.
Microsoft will be ending support for Windows 10 in October 2025. In alignment with this change, we will also conclude official support for VG applications running on Windows 10 starting with version 2025.3.
We encourage affected users to upgrade their operating system to continue benefiting from the full capabilities of VGSTUDIO, VGSTUDIO MAX, VGMETROLOGY, VGinLINE, and other VG software products.
This is an important step towards ensuring our VG software suite operates optimally on modern platforms, providing you with enhanced performance, security, and compatibility.
For assistance with upgrading or any questions, please contact our support team.
With the launch of our new application, VGTRAINER, you now have the power to train your own deep-learning models for segmentation in-house — without the need for AI or programming expertise. Apply these models with the Deep Segmentation module to leverage AI to tackle segmentation challenges previously unsolvable due to CT data quality, while maintaining total control over your confidential part data.
Watch the video tutorial here.
Read the basic requirements below or download the PDF for the complete system requirements for VGTRAINER.
Windows 10* Enterprise 64-bit
Windows 10* Professional 64-bit
Windows 11 Enterprise 64-bit
Windows 11 Professional 64-bit
*VGTRAINER will no longer support Windows 10 as of version 2025.3
Minimum: x86-64 CPU. ARM processors are not supported.
Recommended: Performant Intel or AMD multi-core processors, e.g., AMD Ryzen™ 9 5950X 16-core processor or Intel® Xeon® E5-2687W v3 processor, with 3 GHz or higher
Training a model with VGTRAINER requires a minimum of 128 GB RAM. The actually required size of the main memory strongly depends on the size of your task.
