Over the past few decades, extensive efforts have been dedicated to developing various methods for training supervised pedestrian detectors. The success of these methods largely relies on the availability of large-scale datasets. Unlike generic object detection datasets, several datasets specifically designed for pedestrian detection have been accumulated over the years. These include MIT, INRIA, ETH, USC, TUD-Brussels, and Daimler. Furthermore, some datasets, such as Caltech, CityPersons, and ECP, are also compared in this context.
ORKG Comparisons have changed. We have added new features and improved the user interface. Comparisons might look slightly different, but the comparison data itself remains unchanged.