Cornell PhD student Sean Bell (with Cornell professor Kavita Bala, and Microsoft researchers Larry Zitnick and Ross Girshick) won Best Student Entry in the Microsoft COCO 2015 Detection Challenge. The COCO (Common Objects in Context) competition is a computer vision challenge designed to advance the state of the art in object detection, and is significantly more difficult than previous challenges on object detection. The team developed ION, the Inside Outside Net, that uses deep learning to detect objects in images from the Microsoft COCO dataset. The ION team placed third overall, with first place going to Microsoft Asia, and second place going to Facebook AI research. Results will be presented at ICCV 2015.

Competition results: http://mscoco.org/dataset/#detections-challenge2015

Team: Sean Bell, Larry Zitnick, Kavita Bala, Ross Girshick 
PDFhttp://arxiv.org/pdf/1512.04143.pdf 
Title: Inside-Outside Net: Detecting Objects in Context with Skip Pooling and Recurrent Neural Networks