Communications of the ACM
Genius of Einstein, Fourier Key to New Humanlike Computer Vision
Purdue University researchers have developed two new techniques for computer-vision technology that mimic how humans perceive three-dimensional (3D) shapes by instantly recognizing objects no matter how they are twisted or bent. The techniques, called heat mapping and heat distribution, apply mathematical methods to enable machines to perceive 3D objects, says Purdue professor Karthik Ramani.
Heat mapping works by simulating how heat flows over an object while revealing its structure and distinguishing unique points needed for segmentation by computing the heat mean signature, which enables a computer to determine the center of each segment to define the overall shape of the object. In temperature distribution, heat flow is used to develop a histogram of the entire object. "Albert Einstein made contributions to diffusion, and 18th century physicist Jean Baptiste Joseph Fourier developed Fourier's law, used to derive the heat equation," Ramani says.
The systems simulate heat flowing from one point to another and in the process characterize the shape of an object, he says. The researchers tested the techniques on different complex shapes, and the heat-mapping method enabled the computer to recognize the objects no matter how they were bent or twisted by ignoring noise introduced by imperfect laser scanning and other nonsensical data.
From Purdue University News
View Full Article
No entries found