For technical contributions to the problem of object detection in images which have had very high impact in the fields of computer vision and machine learning.
Pedro Felzenszwalb developed new methods for detecting objects in pictures and video. These methods are key building blocks for most current solutions to object recognition, and have changed the shape of the field. Beginning with his MS and PhD work at MIT, Felzenszwalb has made continuous contributions to the field of object recognition, including highly respected work with D.P. Huttenlocher on the efficient matching of deformable, part-based models. The results of Felzenszwalb's work on object detection with R.B. Girshick, D. McAllester, and D. Ramanan, are the current state of the art. These new methods use a sliding window that is moved around the image, testing the underlying image pattern. Scores that evaluate matches to a set of local patterns, and check whether the local patterns are in about the right place, are combined. The innovative approach, which collects and combines a variety of observations from earlier papers, won the international PASCAL Visual Object Classes Challenge in 2008 and 2009. In addition to his novel algorithms and theoretical results, Felzenszwalb has contributed widely-used, open-source software for computer vision, thereby stimulating new research and applications.Scroll Up
Pedro Felipe Felzenszwalb is the recipient of the 2013 Grace Murray Hopper Award for contributions to object recognition in pictures and video. Felzenszwalb developed innovative methods that have become key building blocks for most solutions to object recognition. His recent approach uses a sliding window that is moved around the image, testing the underlying image data to determine if local patterns are properly located. He also contributed widely- used, open-source software for computer vision, stimulating new research and applications. Felzenszwalb is an associate professor of Engineering and Computer Science at Brown University.