An easy-to-use AI toolkit and SDK for computer vision, suitable for model training and inference.
- Zero-code machine-learning toolkit
- ASUS UX design: intuitive, simple, and quick project setup in four steps.
- Diverse models (object detection, classification, segmentation, and anomaly detection) for vertical applications, and integration of various intuitive labeling tools.
- Unique AI framework supporting Intel® OpenVINO architecture for inference.
- A report wizard for sample filtering, with unique model training and validation.
- Optimized performance for handling simultaneous training projects.
- Language support for Traditional Chinese, Simple Chinese and English
- OS Support for 64-bit edition of Windows 10 and Windows 11 Professional.
- API support for C, C++ and C# to ease integration into existing programs.
As well as the Training mode, the toolkit also includes Runtime mode as an inference engine that supports C, C# and C++ APIs to integrate existing graphical interfaces. They can also retrieve in-line data for analysis and export it for further analytics, visualization, database management, and other related tasks in the edge or cloud. As training can consume a large amount of time, the Scheduler mode can train pre-defined tasks in batches for higher efficiency.
AISVision Project Manager
Managing multiple AI models and use cases
Intuitive labeling tool
Labeling images for AI processing
AISVision integrates various labeling tools for applying important annotations to images for AI Model training. Simply drag the mouse cursor and select the desired label shape, which supports supervised learning, including Classification, Object Detection and Segmentation in AISVision.
Unique validation feature
Simulating time, verifying models and tuning functions
AISVision supports innovative production-line inspection with time simulation to accelerate automated production planning. A unique model adjustment function can quickly adjust the model without requiring retraining. AISVision also provides a Model validation report for reviewing results, helping developers to adjust the training dataset and preview the model with corresponding sensitivity adjustments.
High flexibility inference architecture
Diverse API support and a highly flexible inference architecture
CPU inference with Intel® OpenVINO framework
Special algorithms allow inference with Intel CPUs to achieve the accuracy target by incorporating the GPUs, which increases inference flexibility, lowers energy consumption and reduces the cost of hardware.