Before I start this week’s update, I would like to thank one of our backers, Dan. From an email query they sent, I noticed a mistake in our project’s description and corrected it with Crowd Supply (details below). Thank you, Dan!

This week’s news is that our Software Open Source for EyeCloudAI CDK is updated on GitHub.

Open Source Assets for EyeCloudAI CDK

As you can see in our GitHub repo, our open source assets mainly consist of two parts: hardware and software.

In the "Hardware" folder, you will find all the schematics for all five boards, namely, the core board, the eMMC board, and the three sensor boards.

Clicking on "Software" from the main page will direct you to the OpenNCC project, which is the base of EyeCloud’s Open Software Sources.

In the latest version of OpenNCC(R21.07.02), we restructured the software resources and added scripts for different platforms (Linux, RaspberryPi, Windows, etc.), so developers can easily start from the hardware platforms they are already using and they can check their environment setups with scripts we provide in the CDK.

Importantly, I need to declare that the firmware of EyeCloudAI CDK is not open source, due to the license limitations of Movidius. (This is what Dan had asked about in their email.) The firmware included in the SDK will appear as deployable .mvcmd files.

Compatibility with alwaysAI and PaddlePaddle

Besides OpenVINO, EyeCloudAI CDK also supports AlwaysAI(a powerful, deep-learning computer vision platform) and PaddlePaddle (an R&D deep-learning platform). In the latest SDK, you’ll find three Linkage_demo files under "Examples:"

Below is a demo of an OCR application we developed with PaddlePaddle which is used by NIU electric mopeds for rolling code recognition. Take a look!

That’s all for today. Stay cool and see you next update!