%%%%% Description of vehicle.mat file, the dataset used for "Edge-based %%%%% %%%%% rich representation for vehicle classification" %%%%% Four variables exist in vehicle.mat workspace. train_carvan -> training set for car vs. minivan classification (50 cars, 50 minivans) test_carvan -> test set for car vs. minivan classification (200 cars, 200 minivans) train_sedantaxi -> training set for sedan vs. taxi classification (50 cars, 50 minivans) test_sedantaxi -> test set for sedan vs. taxi classification (200 sedans, 130 taxis) Each variable is a structure array. Fields of the strucuture are: lbl -> Class label (car vs. minivan: car=0, van=1; sedan vs. taxi: sedan=0, taxi=1) im -> Object image mask -> Object mask to separate a moving object from background bdgbox -> Object bounding box computed by an object's contour