Very excited that our most recent paper on #ExplainableAI & #ActiveLearning has been accepted to the #SENTIS workshop at #semanticsconf by @semantics

https://semsys.ai.wu.ac.at/sentis2025/

SENTIS Workshop

Since #Unity seems to have discontinued development of their Barracuda inference engine, I have revisited my infamous YoloHolo sample to make it work with the newer #Sentis engine. Also, the code is now fully cross platform between hashtag#MagicLeap2 and hashtag#HoloLens2 and runs on both without changes, which is, given the message Microsoft telegraphed about the future of HoloLens 2 - or rather, the lack thereof - even more important than it was before. #MixedRealityo
https://localjoost.github.io/Cross-platform-Yolo-object-detection-with-Sentis-on-Magic-Leap-2-and-HoloLens-2/
Cross-platform Yolo object detection with Sentis on Magic Leap 2 and HoloLens 2

Some time ago, I wrote about Yolo object recognition using HoloLens 2 and the Unity Barracuda inference engine. However, Unity seems to have stopped developing this engine and is now pushing the Sentis inference engine. On the same vein: Microsoft have announced they have stopped producing HoloLens 2, with no successor being announced. I have been banging the cross-platform drum for quite a while now, and I think this proves me right.

DotNetByExample - The Next Generation

Ray Traced Pong starting to looking pretty good on integrated graphics (neural rendering using @unity #Sentis)

Caustics, refraction and shadows showcased here.

#MachineLearning #unity #madewithunity