We captured a video sequence datasets for public use, which contain images taken with a Phantom 4® drone, with an HD camera under windy conditions, which affected the image stabilizer capabilities.  Images were resized to 640 x 364 pixels, containing 41,040 to 56,975 ROI annotated objects, with a minimal height size set to 25 pixels. 60% of the annotated data corresponds to occluded motorcycles. Objects partially occluded with height less than 25 pixels were not annotated.

The dataset was annotated by means of Viper annotation tool. The ground truth generated is specified in a XML file which describes the class, frames covered by the object, Name, Id, height and and width of the bbox surrounding the object. For an overview of these datasets please see the table below, for more details or to download a dataset please follow the dataset links.

Dataset Classes Objects Minimal Vertical size Oclussion Level Format
MotorBikes7,500 1 221 25 > 60% XML (Viper file)
MotorBikes10,000 1 317 25 > 60% XML (Viper file)


Original Dataset Image vs Annotated Image.

MotorBikes is a dataset of 7500 anotated images, for download and more details please click here.


Note how the dataset subtly change the perspective from image captures, and also lighting conditions
Now Available!!!


For any queries related to these datasets please email Jorge Espinosa