Hi, I've been looking around to true CAD model recognition, and needless to say, Vuforia seems a very nice option, but they are less than opaque about their pricing, and what is clear is that it's way way way too expensive for small developers. It seems to be geared towards multi-billion dollar industries as far as I can tell. Wikitude has now something similar in beta which I'm checking out, but I'm afraid this also will be on the pricy side. We're set to create AR experiences for Non-profits (museums, cultural enities) How I understand it object recognition is based on matching pointcloud data, and model recognition is a more advanced way of detection, that involves machine learning and based on geometry. Please correct me if i'm wrong Are there similar features in the pipeline for ARFoundation/ARKit/ARCore?
Currently, in AR Foundation's environment, only ARKit is capable to recognize objects previously scanned with a scanning app.
Thanks. Yes I understand. My question is about future developments in 3D model recognition (and my assumption about the difference between object recognition and model (CAD) recognition as seen in Vuforia) cheers
There are various services that offer 3D model recognition and tracking (e.g. VisionLib). AR Foundation is extensible, so if you find a provider that meets your needs you can ask them to integrate with AR Foundation.