phyBOARD^®-Pollux AI kit

Description:

The phyCORE-i.MX 8M Plus AI Kit serves both as a learning platform and as a reference design to unlock the embedded vision and machine learning functionalities of the i.MX 8M Plus processor. 

A cost-effective and industry-standard VM-016 HD camera with MIPI CSI-2 interface is included in the kit for capturing the image data. The data is processed by the internal Image Signal Processor of the i.MX 8M Plus, which means that post-processing of the image data is not necessary. 

The development environment for NXP's eIQ machine learning software is already in the BSP integrated. With the support of common models such as pytorch, TensorFlow Lite and ONNX format, you can start directly with your application. Hardware specially designed for machine learning or other tools for using neural networks are not required.

The data processing takes place exclusively at the edge, which guarantees an additional time advantage.
With the phyCORE-i.MX 8M Plus, embedded designs can now be used for AI and machine learning applications that require high-performance data processing for the rapid implementation of AI processes.

Face recognition demo application

Using the example of a demo application, it is demonstrated how neural networks can be used to compare objects using machine and deep learning algorithms. 
Specific facial features are extracted from recorded live images of faces by various neural networks, which are then compared with pre-trained faces of prominent people. The face with the smallest deviations from your own face is displayed as the best hit.