Ross Girshick is a computer vision researcher best known for the R-CNN family of object-detection methods. He holds a Ph.D. from the University of Chicago and was a postdoc at UC Berkeley before working at Microsoft Research, Facebook AI Research (FAIR), and the Allen Institute for AI.
In 2013 Girshick invented R-CNN (“Regions with CNN features”), which carried the deep-learning revolution from image classification into detection by running a convolutional network over candidate object regions. He then drove a rapid sequence of improvements - Fast R-CNN and Faster R-CNN - that made detection far faster and more accurate, and he is the author of the widely used Detectron open-source detection library. He also contributed to later landmarks including Mask R-CNN and Segment Anything. His detection work has drawn hundreds of thousands of citations.
His contributions have been recognized with the PAMI Young Researcher Award, three Mark Everingham Prizes for open-source and dataset work, and multiple test-of-time awards, including the 2024 Longuet-Higgins Prize for R-CNN. In late 2024 he co-founded the AI startup Vercept.