The following are sample representative images from our LISA-Face Detection Database with overlaid ground truth annotations. All images in our database are taken from videos of looking at the occupants inside the vehicle while driving on highways and urban streets. Some of the original video data has been captured on our test beds, while some are kindly provided by YouTube. These images from videos are gathered for research purposes only and we do not endorse any messages from respective videos. Note that only a small representative sample of the actual LISA Face Detection Database is shown here to illustrate the challenges faced when looking at driving data, such as illumination, occlusion, motion blur, camera perspective and image quality. The actual test set will be released soon. Interested participants, please contact us and we will keep you updated. Let the challenge begin!