Geek asks and answers

How does Vuforia image recognition work?

I use the Vuforia SDK to create an Android app, and I wonder how marker tracking works. Does the application convert video frames to byte codes, and then compare them with a .dat file created by creating a marker? Also, where is this code located in the Vuforia sample application, is it in C ++? Thank.

+5
source share
3 answers

Well, you do not see the code for recognition and tracking, as they are Qualcomm's intellectual property and usually should not be disclosed. Vuforiais not an open source library.

Vuforia first detects the “function points” in your target image [web target control], and then uses the data to compare the functions in the target image and the receiving frame from the camera.

Google's “Natural Detection and Tracking Function,” which falls under the “Computer Vision” area, and you'll find interesting content.

+5
source

No, the detection and tracking code posted in libQCAR.so But the question “how does it work” lies in the complex answer here. If you want to be familiar with the detection and tracking of objects - start learning the method like mser, surf, sift, ferns and others.

+2
source

Vuforia . , Vuforia. , , SIFT.

+1

More articles:


All Articles