Deploy Vision AI into your production systems with just a few mouse clicks. Inference provides a comprehensive and configurable execution environment for Deep Neural Networks. Through a user-friendly interface, trained AI models can be quickly and easily executed for detection, classification and anomaly detection tasks. The results can be accessed via standard interfaces and displayed via the Inference user interface. This process takes only a few minutes. Time-consuming interface developments are now a matter of the past.
The runtime environment of Data Spree Inference makes it possible for you to quickly and easily deploy different AI models for various applications in your facility. Inference supports a variety of AI model types for Image Classification, RoI Classification and Object Detection. Even the execution of different AI models at the same time (Multi-Model Inference) is possible without any additional effort.
In the well-organized user interface of Data Spree Inference, AI models can be easily selected, integrated and executed within just a few mouse clicks. Programming is thus no longer necessary. Live data from cameras and sensors can be displayed, recorded and saved easily via the user interface. From now on, you can access the execution results of your system at any time via the live view.
The wide-ranging and vendor-independent interfaces make connecting different camera and sensor hardware easy and user-friendly. AI models based on different camera and sensor data, whether 2D images, 3D point clouds, false color images or scanner data can be easily executed, displayed, captured and stored with Data Spree Inference.
Design your own AI runtime environment. Through the Inference user interface, you can compose, adjust and manage different AI networks based on different cameras and sensors in a clear and modular way. Monitor the execution and results in real time along the entire process pipeline.
The open plugin structure of Data Spree Inference opens all possibilities for customizing your solution. Data Spree already provides many ready-to-use plugins for various applications, such as object tracking and counting. Integrate your own Python-based plugins for your individual needs.
The rollout of AI models into the production environment requires the right hardware with the appropriate computing capacity. The selection of the optimal hardware and computing system depends on your requirements in terms of data quantity, execution time and extra computational operations. There are numerous options for deploying AI models with Data Spree Inference: from smart cameras to industrial PCs to server racks, everything is possible. We support x86 and ARM-based platforms such as NVIDIA® Jetson™ and Intel® Movidius™.