EBDD: EEE BUET Distracted Driving - Video Database
Introduction:

Road accident is one of the ten leading causes of death all over the world. Distraction during driving is one of the major causes of road accidents. Cautious driving is intensely becoming an apprehension for global road safety in recent years. For this reason, elimination of distraction has emerged as a significant area of research in the intelligent transportation system. Distraction during driving can occur in many forms. Typical distractions can be talking or texting on cell phone, operating cabin equipments, or eating food. If these common forms of distractions are identified autonomously, then the drivers can be warned to bring back their focus on the road. Thus, the possibility of accidents due to distracted driving can be greatly reduced.

For autonomous identification of distractions, algorithms have to be developed and their performances have to be tested on a database. A sufficient number of quality video clips that depict distracted driving cannot be found in the World Wide Web. In particular, the clips found do not depict distraction as a whole or are constrained to brief lengths or inconsistent capturing angles. Owing to these limitations, experiments cannot be performed on the video clips available in the web. Hence, for facilitating research on autonomous identification of distractions during driving, the EEE BUET Distracted Driving (EBDD) video database has been developed. The clips of the database depict cautious driving as well as activities that cognitively distract drivers during driving. Any researcher reporting results which use this database must acknowledge the EBDD database. It is required that the researchers do so by citing the publication:

  • Tashrif Billah, S. M. Mahbubur Rahman, M. Omair Ahmad, and M. N. S. Swamy, “Recognizing distractions for assistive driving by tracking body parts,” in-press, IEEE Transactions on Circuits and Systems for Video Technology, DOI: 10.1109/TCSVT.2018.2818407, pp. 1-15, Mar. 2018.

@ Article{EBDD_Database, 
author =   "Tashrif Billah, S. M. Mahbubur Rahman, M. Omair Ahmad, and M. N. S. Swamy", 
title =    " Recognizing distractions for assistive driving by tracking body parts",
journal =  "IEEE Transactions on Circuits and Systems for Video Technology", 
DOI =   "10.1109/TCSVT.2018.2818407",
pages =  "1-15", 
year = 2018}

Early access available: http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=8322191

In addition, we appreciate to hear about any publications that use the EBDD database. Feedbacks on the database and this website are also welcome. We are continuously updating the database.

Data Description:
A good resolution camera (Sony Cyber Shot 14.1 mega pixels) was affixed on the front windshield facing the driver inside the vehicle. Videos of a number of drivers were captured in daylight on city roads and university campus in Dhaka, Bangladesh. The formed data set is diverse in terms of landscape, illumination, vehicle, or road condition (smooth or bumpy). The age and experience of the drivers also vary significantly. The videos in the database can be broadly categorized as cautious or distracted driving. The distracted driving can be of four types - talking on cell phone, eating, texting on cell phone, and operating cabin equipments (unattentive). The videos available here can be used for activity recognition of a driver. All the videos have frame size of 854x480 pixels with a frame rate of 30 fps. The characteristics of videos of cautious and distracted driving are described in Tables 1 and 2, respectively, and typical frames of distractions are shown in Figure 1.

 
Figure 1: Sample scenes of common distractions during driving
Table 1: Characteristics of video clips with cautious driving

Clip ID (.mp4)

Duration (min:sec)

Characteristics

Size (MB)

Experience of Driver

Viewing Angle

Illumination

AREFIN_ATTENTIVE

2:00

Professional

Front

Normal

72.9

EHSAN_ATTENTIVE

1:59

Amateur

Front

Low

28.4

JEWEL_ATTENTIVE

2:00

Professional

Front

Low

32

NAVID_ATTENTIVE

1:59

Amateur

Front

Normal+Sunny

40.6

SOHEL_ATTENTIVE

1:58

Professional

Front

Low+Sunny

33.7

SUMON_ATTENTIVE

2:00

Professional

Front

Normal+Sunny

52.4

TASHRIF_ATTENTIVE

1:59

Amateur

Front

Normal

34.6
Table 2: Characteristics of video clips with distracted driving

Clip ID (.mp4)

Duration (min:sec)

Characteristics

Size (MB)

Experience of Driver

Viewing Angle

Illumination

AKKAS_CELL_PHONE

00:29

Professional

Front

Rainy

9.37

AKKAS_EATING

00:30

Professional

Front

Rainy

14.66

AKKAS _TEXT

00:32

Professional

Front

Rainy

14.91

AKKAS _UNATTENTIVE

00:29

Professional

Front

Rainy

12.67

AKTAR_CELL_PHONE

00:36

Professional

Front

Low

8.20

AKTAR_EATING

00:30

Professional

Front

Low

8.98

AKTAR_TEXT

00:31

Professional

Front

Low

7.22

AKTAR_UNATTENTIVE

00:36

Professional

Front

Low

9.19

ALAMIN_CELL_PHONE

00:31

Professional

Front

Rainy

16.15

ALAMIN_EATING

00:32

Professional

Front

Rainy

15.10

ALAMIN_TEXT

00:29

Professional

Front

Rainy

12.49

ALAMIN_UNATTENTIVE

00:30

Professional

Front

Rainy

15.09

AREFIN_CELL_PHONE

00:29

Professional

Front

Normal

11.52

AREFIN_EATING

00:38

Professional

Front

Normal

32.6

AREFIN_TEXT

00:34

Professional

Front

Normal

22.23

AREFIN_UNATTENTIVE

00:26

Professional

Front

Normal

22.66

ASIF_CELL_PHONE

00:30

Amateur

Front

Normal+Low

17.82

ASIF_EATING

00:29

Amateur

Front

Normal+Low

13.82

ASIF_TEXT

00:28

Amateur

Front

Normal+Low

13.48

ASIF_UNATTENTIVE

00:35

Amateur

Front

Normal+Low

19.07

EHSAN_CELL_PHONE

00:33

Amateur

Front

Rainy

13.12

EHSAN_EATING

00:31

Amateur

Front

Rainy

12.09

EHSAN_TEXT

00:26

Amateur

Front

Rainy

11.05

EHSAN_UNATTENTIVE

00:44

Amateur

Front

Rainy

15.03

JEWEL_CELL_PHONE

00:33

Professional

Front

Rainy

14.19

JEWEL _EATING

00:32

Professional

Front

Rainy

15.17

JEWEL _TEXT

00:33

Professional

Front

Rainy

16.05

JEWEL _UNATTENTIVE

00:36

Professional

Front

Rainy

14.99

NAVID_CELL_PHONE

00:29

Amateur

Front

Rainy

13.24

NAVID _EATING

00:29

Amateur

Front

Rainy

11.25

NAVID _TEXT

00:33

Amateur

Front

Rainy

12.79

NAVID _UNATTENTIVE

00:28

Amateur

Front

Rainy

11.93

RIPON_CELL_PHONE

00:30

Professional

Front

Normal+Sunny

10.29

RIPON_EATING

00:31

Professional

Front

Normal+Sunny

15.71

RIPON_TEXT

00:31

Professional

Front

Normal+Sunny

13.54

RIPON_UNATTENTIVE

00:30

Professional

Front

Normal+Sunny

14.77

SOHEL_CELL_PHONE           

00:30

Professional

Front

Normal+Low

10.49

SOHEL _EATING

00:30

Professional

Front

Normal+Low

16.39

SOHEL _TEXT

00:30

Professional

Front

Normal+Low

15.99

SOHEL _UNATTENTIVE

00:32

Professional

Front

Normal+Low

17.42

SUMON_CELL_PHONE

00:30

Professional

Front

Rainy

10.18

SUMON _EATING

00:30

Professional

Front

Rainy

12.99

SUMON _TEXT

00:29

Professional

Front

Rainy

10.56

SUMON _UNATTENTIVE

00:30

Professional

Front

Rainy

10.38

TASHRIF_CELL_PHONE        

00:27

Amateur

Front

Normal

13.61

TASHRIF _EATING

00:29

Amateur

Front

Normal

11.21

TASHRIF _TEXT

00:33

Amateur

Front

Normal

14.97

TASHRIF _UNATTENTIVE

00:31

Amateur

Front

Normal

14.41

ZAKARIA_CELL_PHONE

00:29

Amateur

Front

Normal+Sunny

18.79

ZAKARIA _EATING

00:29

Amateur

Front

Normal+Sunny

16.52

ZAKARIA _TEXT

00:39

Amateur

Front

Normal+Sunny

17.60

ZAKARIA _UNATTENTIVE

00:29

Amateur

Front

Normal+Sunny

10.01
Using EBDD Database:
The database is downloadable free of cost but its use is limited for academic or research purpose only. However, any research outcome using the database is required to cite the name and address of EBDD Video Database. Any reproduction, redistribution or tampering of the database in part or as a whole is prohibited. Typical samples of tracking videos can be found here.
Download:
The EBDD video database is available here for download. The file of the database is a zip file of 1.02 GB that contains a folder named EBDD_Video_Database. The password to unzip the file can be obtained on request to mahbubur@eee.buet.ac.bd or mahbuburbd@gmail.com. We also warmly welcome any advice regarding the improvement of the quality and content of the database.

Your life matters to your family and loved ones - drive safe!