Data Set Information:

DrivFace数据库包含真实场景中驾驶时的主题图像序列。它由606个样本组成，每个样本为640-480像素，在不同的日子里从4名司机（2名女性和2名男性）那里采集，这些司机具有眼镜和胡须等多种面部特征。

Additional files:
drivFace.mat contains the dataset in Matlab (under prtools library) with the driver faces normalised to 80x80 pixels each and their associated gaze direction labels a€?looking-righta€?, a€?frontala€? and  a€?looking-lefta€?.

Attribute Information:

The ground truth contains the annotation of the face bounding box and the facial key points (eyes, nose and mouth).
A set of labels assigning each image into 3 possible gaze direction classes are given.
The first class is the a€?looking-righta€? class and contains the head angles between -45?o and -30?o.
The second one is the a€?frontala€? class and contains the head angles between -15?o and 15?o.
The last one is the a€?looking-lefta€? class and contains the head angles between 30?o and 45?o.

Files and scripts
a€￠ DrivImages.zip has the driver images. The imag's name has the format:
   * YearMonthDay_subject_Driv_imNum_HeadPose.jpg

i.e. 20130529_01_Driv_011_f .jpg is a frame of the fisrts driver corresponding to the 11 sequence's image and the head pose is frontal.
subject = [1:4], imNum = [001:...], HeadPose = lr (looking-right), f (frontal) and lf (looking-left).
 

a€￠ drivPoints.txt contains the ground truth in table's format, where the columns have the follow information:
   * fileName is the imagen's name into DrivImages.zip
   * subject = [1:4]
   * imgNum  = int
   * label   = [1/2/3] (head pose class that corresponding to [lr/f/lf], respectively)
   * ang     = [-45, -30/ -15 0 15/ 30 15] (head pose angle)
   * [xF yF wF hF] = face position
   * [xRE yRE]     = rigth eye position
   * [xLE yL]      = left eye position
   * [xN yN]       = Nose position
   * [xRM yRM]     = rigth corner of mouth
   * [xLM yLM]     = left corner of mouth  


a€￠ read_drivPoints.m is a Matlab function to read the drivPoints file. You can also use:
   * Table = readtable('drivPoints.txt');

a€￠ drivFace.mat contains the dataset in Matlab (under prtools library) with the driver faces normalised to 80x80 pixels each and their associated gaze direction labels a€?looking-righta€?, a€?frontala€? and  a€?looking-lefta€?.

Relevant Papers:

Katerine Diaz-Chito, Aura Hern??ndez-Sabat??, Antonio M. L?3pez, A reduced feature set for driver head pose estimation, Applied Soft Computing, Volume 45, August 2016, Pages 98-107, ISSN 1568-4946,

Citation Request:

Katerine Diaz-Chito, Aura Hern??ndez-Sabat??, Antonio M. L?3pez, A reduced feature set for driver head pose estimation, Applied Soft Computing, Volume 45, August 2016, Pages 98-107, ISSN 1568-4946,

Katerine Diaz-Chito*
Aura Hern??ndez-Sabat??
Antonio M. L?3pez

Centre de Visi?3 per Computador
Universitat Aut?3noma de Barcelona

*Corresponding author:{kdiaz '@' cvc.uab.es}