SEAS-FR-DB



Welcome

Hi, welcome to the SEAS-FR-DB!

Face Recognition Database for Videos with 1080p and 30fps.

You can check More Details by scrolling down!



Description

The database is designed for providing high-quality HD multi-subject benchmarked video inputs for face recognition algorithms. The database is a useful input for offline as well as online(Real-Time) Video scenarios. The database has primarily three classified subjects (3) and two not-classified(unknown/un-labeled) subjects available in 30fps - High Definition Video (Full HD - 1080p) video.

Hope you love this database. We would love to hear your inputs. Share your amazing work with us!

If you use this database, or refer to its results, please cite us as per following:

Jobanputra, Mayank, Axat Chaudhary, Saumil Shah, and Ratnik Gandhi. "Real-time face recognition in HD videos: Algorithms and framework." In Systems Conference (SysCon), 2018 Annual IEEE International, pp. 1-8. IEEE, 2018.


Database

3 Test Subjects

4 Videos for Each Subject

For Each Video, Duration : 4 sec , Total Frames : 120 frames


Videos

To the left is the File Hierarchy of Videos in our Database. There are 2 sub directories Training Set and Testing Set. Training Set has 4 sub-folders s1 to s3 each having videos of subjects 1 to 3 respectively. Testing Set has 2 videos that are different from the Training Set.
For any video file starting with 's....mov',

Format : sX_V.mov ; Where, X = Subject No & V = Video No

Example: For Subject - 3, Video-2:
Therefore, X = 3,V = 2, F = S, L = S, Z = 1 - 120

File Names would be : "s3_2.mov"

Which is at Location - "SEAS-FR-DB/Videos/Training Set/s3/s3_2.mov"











Training Images

To the left is the File Hierarchy of Training Images in our Database. There are 3 sub directories for each Subject. Each directory has 4 other sub-folders each having extracted images of the corresponding video of that subject from the Training Set Videos.
For any image file starting with 's... .jpg',

Here, Format : sX__FL_V_Z.jpg ;

Where,
X = Subject No
F = Subject's First Name Initial
L = Subject's Last Name Initial
V = Video No (Video from Testing Set Directory)
Z = Frame No

Example: For Subject - 3, Video-2:
Subject Name is : Saumil Shah
Therefore, X = 3,V = 2, F = S, L = S, Z = 1 - 120

File Names would Range from,
"s3__SS_2_1.jpg" to "s3__SS_2_120.jpg"

Training Video is at Location - "SEAS-FR-DB/Videos/Training Set/s3/s3_2.mov"







Subjects

3 Trained Subjects that act as Known Subjects

2 Subjects that act as Unknown Subjects

Subject1 Subject2 Subject3

Subject4 Subject5



Contributors

Saumil Shah

Axat Chaudhary

Mayank Jobanputra



Acknowledgements

We sincerely thank Dr. Ratnik Gandhi, Dr. Mehul S. Raval for their constant guidance and support for our Face Recognition project and SEAS-FR-Database.




Get the Database


Benchmarking Details can also be found at Github.
If you use this database, or refer to its results, please cite us as per following:

Jobanputra, Mayank, Axat Chaudhary, Saumil Shah, and Ratnik Gandhi. "Real-time face recognition in HD videos: Algorithms and framework." In Systems Conference (SysCon), 2018 Annual IEEE International, pp. 1-8. IEEE, 2018.

BibTeX:

@inproceedings{DBLP:conf/syscon/JobanputraCSG18, author = {Mayank Jobanputra and Axat Chaudhary and Saumil Shah and Ratnik Gandhi}, title = {Real-time face recognition in {HD} videos: Algorithms and framework}, booktitle = {2018 Annual {IEEE} International Systems Conference, SysCon 2018, Vancouver, BC, Canada, April 23-26, 2018}, pages = {1--8}, year = {2018}, crossref = {DBLP:conf/syscon/2018}, url = {https://doi.org/10.1109/SYSCON.2018.8369529}, doi = {10.1109/SYSCON.2018.8369529}, timestamp = {Tue, 12 Jun 2018 18:04:41 +0200}, biburl = {https://dblp.org/rec/bib/conf/syscon/JobanputraCSG18}, bibsource = {dblp computer science bibliography, https://dblp.org} }