AI Guru Software Features

Setting the ASRock NUC BOX-155H apart from other Mini-PCs is the inclusion of AI tools. To put it simply ASRock has created an AI tool kit based on seven pre-trained AI models using Intel's Distribution of OpenVINO Toolkit. It offers the ability to utilize the CPU, iGPU, or NPU as the inference device for Face Recognition, Object, Detection, Human Pose Estimation, Image Inpainting, Face Detection, Gaze Estimation, and Interactive Face Detection. ASRock Industrial has a full PDF guide on how the software works and how to use it which makes getting started quite easy.


While I found the features interesting and in some cases quite useful the speed of the output leaves a lot to be desired. This has less to do with ASRock and more to do with the infancy of NPU performance and AI processing performance. This is a Mini-PC after all, so expecting earth-shattering performance is unrealistic, however, what is demonstrated here is an entirely useful setup, that while slow in my opinion, could be useful in the right situations and considering the rapid pace of advancements means as time goes on the overall performance and usefulness of AI Guru, and AI inference equipped software in general will only get better.


That said, considering the subject at hand ASRock's push to include it and deliver a usable experience should be commended. But for now I find the overall performance lacking at least in the tests I could run with my own data sets / photos etc. Still, I can see the draw here, and it gives us a peek at where things are headed. Considering my girlfriends love of photography the image based Face Recognition model will prove especially useful. For now, it's a niche reason to consider the system but unlike ARGB lighting and displays mounted on CPU coolers, the various trained AI models here actually have utility and usefulness. All I can say for now is that while I don't think these features improve my own personal workflow in any way at this time. I am excited for future developments in this particular field.