A Virtual Object Generation Method Based on Image Vector Recognition

A new virtual object generation method for AR applications based on image vector recognition is introduced in this paper. During the whole process, data acquiring and virtual object generation are completed automatically, including image smoothing, image edge detection, target silhouette tracking, vector curve compression and real-time rendering. This method improved the degree of automation of virtual object generation in AR applications rather than using pre-built models with digital content creation software. The key steps of the approach are described in detail and a practical application is also given.