Abstract
AbstractPerceptual grouping is the basic capability of the human brain to derive relevant structure without any prior content knowledge. This phenomenon has been described in the human visual system by Gestalt psychologists, and later, it was recognized by the computer vision community as a method for reducing the computational complexity of visual recognition. The article starts by introducing the laws of Gestalt, and it continues by describing how they are used to obtain perceptual grouping in computer vision. Different algorithms are presented for achieving perceptual grouping in two and three dimensions, at the signal, primitive, structural, and assembly levels. Spatiotemporal grouping is also discussed. The most successful computation techniques are described. Finally, a few methods for performance evaluation are also presented.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
