Annotation of Blurred Cabin Imagery for Passenger Detection in SHRP2 NDS Data

This data set consists of annotations of blurred cabin imagery from two sources: VTT1_OONZ5I This imagery data was originally compiled for Oak Ridge National Laboratory (DOI: 10.15787/VTT1/OONZ5I) . and consists of two images for each vehicle that participated in the SHRP 2 study, usually one from daytime and one from nighttime. This data set had a total of 5,658 blurred cabin images. Annotations include information about the location and type of occupants (driver, front-seat passenger, back-seat passenger). These annotations are stored in json type format following the COCO standard. BCI_TrainingD This imagery data was extracted from a SHRP 2 related data set, generally termed the “SHRP2 Training Data”. This data set has a total of 1,058 blurred cabin images. Annotations include information about the location and type of occupants (driver, front-seat passenger, back-seat passenger). These annotations are provided in .txt files for each image following a Yolo standard (confidence, x_center/image_width, y_center/image_height, box_width/image_width, box_height/image_height).