View Document Preview and Link Document Date: 2014-06-04 23:24:55 Open Document File Size: 529,99 KB Share Result on Facebook
City F. Durand / Paris / / Company Neural Networks / / / Event Product Recall / Product Issues / / Facility University of Coimbra / Jorge Batista Institute of Systems / / IndustryTerm element-wise product / convolution algorithms / dense search / computer vision algorithms / closed-form solutions / dot-product / proposed translation machinery / hard negative mining / mining / augmented real solution / hard negative mining steps / closed-form solution / iterative hard negative mining / negative mining / dot-products / learning algorithms / hard negative mining rounds / / MarketIndex Caltech Pedestrians Detection / Caltech Pedestrian Detection / / Organization North Atlantic Treaty Organization / Cambridge university / University of Coimbra / Jorge Batista Institute of Systems and Robotics / MIT / U.S. Securities and Exchange Commission / / Person B. Alexe / V / Rui Caseiro / / Product rounds / Recall Recall / 0.5 Precision 1 0 0.5 Precision / 0.8 / / ProgrammingLanguage MATLAB / DC / / PublishedMedium Journal of Machine Learning Research / / Technology neuroscience / convolution algorithms / computer vision algorithms / 2 Algorithms / training algorithm / http / Machine Learning / / URL www.isr.uc.pt/∼henriques / / SocialTag 
Beyond Hard Negative Mining: Efficient Detector Learning via Block-Circulant Decomposition Jo˜ao F. Henriques, Jo˜ao Carreira, Rui Caseiro and Jorge Batista Institute of Systems and Robotics, University of Coimbra {hen