PreAct has a powerful Computer Vision platform for creating AI models to detect and understand human gestures and movements.
The Creator leverages the latest image recognition features to generate AI models that can classify a variety of human behaviors. Our end-to-end platform lets you create custom detection engines that are both high-performance and robust.
The Creator can be used to train detection models for any human activity, gesture or movement, including any combination of person and object.
Leveraging both semi-supervised and incremental learning techniques, The Creator generates models that are more accurate and robust with far lower training data requirements.
Light run-time detection engine runs locally on edge devices, and can take advantage of GPUs/TPUs if available. Multiple detection actions can be combined into a single engine for maximum computing efficiency.
The Creator can be accessed via the customer user company's cloud services account, assuring complete confidentiality of all the company training data and models as it's always in-house.
Use real-world or synthetic images and videos to create the dataset needed to train a detection engine. Label your dataset using our built-in annotation tool, or import your annotated data elsewhere.
Need more information about the supported data format?
Choose among our base model that best fits your needs, configure your training parameters and understand the results, with no machine learning expertise.
Need more information about the supported data format?
Export your detection engine from The Creator and reduce development effort by integrating the trained model to your application using the SDK.
