DAIMLER PEDESTRIAN DETCTION BENCHMARK As part of our publication M. Enzweiler and D. M. Gavrila. “Monocular Pedestrian Detection: Survey and Experiments”. IEEE Trans. on Pattern Analysis and Machine Intelligence (available online: IEEE Computer Society Digital Library, http://doi.ieeecomputersociety.org/10.1109/TPAMI.2008.260), 17 Oct. 2008. we are making a large image data set available, to allow benchmarking of pedestrian (object) detection algorithms. The so-called Daimler benchmark * involves a specified training and test set. The training set contains 15.560 pedestrian and non-pedestrian samples (image cut-outs) and 6744 additional full images not containing pedestrians for bootstrapping. The test set involves a sequence with more than 21.790 images with 56.492 pedestrian labels (fully visible or partially occluded), captured from a vehicle during a 27 min drive through urban traffic, at VGA resolution (640x480). As such, the dataset is realistic and one order of magnitude larger than other datasets at time of publication (8.5 Gb). * specifies two evaluation settings: one “generic” (2D bounding box overlap criterion) and one specific to pedestrian detection onboard a vehicle (3D localization criterion, known ground plane and sensor coverage area provide regions of interest, processing constraints). * specifies performance metrics both at the frame- and trajectory-level (the latter also allows benchmarking of tracking algorithms). * provides the baseline performance of three state-of-the-art methods (wavelet-based AdaBoost cascade, HOG/linSVM and a convolutional network NN/LRF) on the specified training and test set. * is made freely available for non-commercial purposes (both training and test set). The Daimler pedestrian detection benchmark is available at: http://www.science.uva.nl/research/isla/downloads/pedestrians/index.html For more information on the pedestrian detection application, see http://www.gavrila.net/Research/Pedestrian_Detection/pedestrian_detection.html Feedback is welcome. ------------------------------------------------------------------ Prof. Dr. Dariu M. Gavrila Environment Perception Daimler Research & Development, Wilhelm Runge St. 11, 89081 Ulm, Germany Email: [first-name].[last-name]@daimler.com - Intelligent Systems Lab Amsterdam Faculty of Science, Room F0.41 University of Amsterdam Science Park 107, 1098 XG, Amsterdam, The Netherlands Email: d.m.[last-name]@uva.nl Web: www.gavrila.net ------------------------------------------------------------------