Algorithm question
Hello, I was wondering if it'd possible to know if MAXST's tracking algorithm is part of Optical Flow algorithms or Feature Descriptor Matching algorithms, since these two classes are very similar: the first one is dense, the second one is sparse. If I had to take a guess, I'd probably say it belongs to the second one, but I can't find any concrete information to confirm my guess, it's just basically a huntch.
Hi,
It's interesing because you're the first person who asks about the algorithm.
We use optical flow and feature descriptor as you already assumed.
Our sdk have image tracking, visual slam and instant tracking for main features and they are based on feature descriptor matching algorithm.
In particular instant tracking, it have visual inertial odometry.
Best regards,
Jack
Maxst SDK team
First of all, thank you for the reply.
Yes, I know it maybe was a strange question to ask, but I needed this information for my Master's Thesis, since I have to write something about the tracking algorithm used by the SDK.
I suppose you can't tell me if it uses the Kalman filter also? It'd be the last thing I ask about this matter, I promise :)
Best regards.
Yes, our sdk uses filter when tracking performace is not good like on featureless environment.
Developers can toggle that filter function via api.
Anyway, we are wondering if you are postgraduate student or not. And you don't need to stop asking us :-)
Jack
Maxst SDK team
Ok, so only Image Tracking uses the Kalman filter or is it implemented in the other SDK available options (Instant Tracker, VSLAM) as well? I'm interested only in the Image Tracking function, since my work is based on this feature.
And yes, I'm a postgraduate student.
Filter toggle option is only for image tracking ^^
Jack
Ok, thank you :)