vAnonymize is a software solution for processing personal data in images and videos. Using artificial intelligence, it recognizes faces and license plates and then anonymizes them. vAnonymize is trained using open-source datasets and offers excellent recognition of all identifiable faces and license plates.
Highlights
- Securely process large amounts of data: Work in compliance with data protection regulations while maintaining high throughput
- Local processing: No internet connection required during use
- No external communication: Data remains within a closed system
- Seamless compatibility: A wide range of data formats is supported
- No data loss: Metadata is preserved during anonymization
Video Recording Under Strict Data Protection Regulations
Camera systems are used in a wide range of applications, including traffic monitoring, web and dashcams, retail, scientific research, healthcare, and industrial manufacturing. Large volumes of video and image data are also generated daily in development projects for advanced driver assistance systems (ADAS) and autonomous driving.
This data is often captured in public spaces, where individuals, vehicles, or license plates may be clearly identifiable. Since this constitutes personal data, applicable data protection regulations, such as the GDPR in the European Union, must be strictly observed. This applies regardless of the specific use case, as soon as identifiable features are recorded.
On-Premise Anonymization
Upload videos and images from any recording system to your local environment and route the data through the anonymization workflow.
Files from an input folder are intelligently distributed across multiple instances for high throughput. Use your existing CPUs and GPUs to speed up the anonymization process. There is only insignificant loss of data quality during the processing, the input format corresponds to the output format. Codec, container and frame rate remain unchanged during anonymization.
The software vAnonymize works completely offline during runtime and without sharing sensitive data with third parties.
Supported Data Formats/Codecs
- AVI (H.264 (Baseline-Profile), YUV420p, MJPEG, YUVj420p, RAW, BGR24/RGB24)
- MP4 (H.264 (Baseline-Profile), YUV420p, MJPEG, YUVj420p)
- JPG (JPEG (ITU T81), JFIF (ITU T871))
- PNG (RGB, RGBA, GRAY, GRAYA, Output: RGB24)
- MF4 (H.264 (Baseline-Profile), YUV420p, JPEG (ITU T81/ITU T.871), YUV422, RAW12)
- Rosbag (JPEG (ITU T81), JFIF (ITU T871))
Product Descriptions
| Component | Requirement | Comment |
| Operating System | Linux with Kernel version >3.10 | Both requirements are necessary to run the tool on a CPU - e.g. for evaluation |
| Hard Drive Capacity (free space) | 60 GByte or more of free disk space | |
| Additional requirements for GPU usage | ||
| Graphics Card | NVIDIA CUDA-capable GPU 12 Gbyte or more of VRAM based on the Turing architecture (7.5 or higher) | |
| Driver | NVIDIA Driver Version ≥570 | |
| Libraries | CUDA Version 12.8 cuDNN Version 9.8.0 | Enables using CUDA as execution provider via CLI command. |
| TensorRT Version 10.9.0.x | Enables using TensorRT as execution provider via CLI command. |
Frequently Asked Questions
Related Products
