ᎣᏏᏲ. ᏙᎯᏧ? Hey, welcome back!
One of the big issues I’ve been working on has been translating OpenCV detection points to BabylonJS coordinates. Up to this point, I’ve had some success getting a marker to track an object.
I decided to take a look at the JSArtoolkit5 examples and see what I could glean from them. The idea was to take the BabylonJS example and apply a live video detection instead of the video loaded. That is successful. Since everything in object detection is basically marker-based detection I could take the code I did for face detection and sending those markers to the JSArtoolkit code. I am not, though.
I found some BabylonJS and BabylonAR code that starts what I’d like to do so I’m making changes that I need to hook more OpenCV code in and then I’m going to publish some sample release files to the GitHub fork I’ve made to test them out.
I’m working for a WebRTC selection to determine which audio or video you wish to use then hook that with OpenCV. That’s tonight’s work. I’m building BabylonJS now and hopefully, I can get it done and uploaded.
Until next time. Dodadagohvi. ᏙᏓᏓᎪᎲᎢ.