Home

Setup

We have built a human face dataset of high-speed videos captured from 8 different (but fixed) viewpoints of faces performing various expressions. We collected recordings of 52 identities in a lab setting. Cameras were positioned on a circular grid, ensuring that the facial area was fully covered. Cameras were synchronized in the sense that we know the correspondence between frames of one camera to another. The length of each recording is from 75 to 90 seconds. Blackfly S BFS-U3-16S2C-CS cameras with a resolution of 1440 by 1080 pixels were used for data acquisition. We additionally added one more Canon DSLR camera.

Collected multi-view videos enable to leverage the temporal correlation between neighbouring frames. An example of the eight multi-view videos is shown below.

Specification

Our dataset contains the following expressions:

• Anger
• Contempt
• Disgust
• Fear
• Joy
• Sad
• Surprise
• Random
• Random with black hat covering hair

Dataset specification:

• Number of identities: 52
• Number of cameras: 8
• Camera Model: PointGrey Blackfly S BFS-U3-16S2C-CS
• Resolution: 1440x1080
• Length of individual recording: 75-90 seconds
• Video Frame-Rate: 112

Apply for Dataset

1. Download and read the license agreement carefully, confirming that all the terms are agreed.
2. Sign the license agreement, scan it and save the scan as a pdf document.

   Electronic signature is allowed, which is turning the handwritten signature into a picture to be placed on the agreement.

3. Send an e-mail to paolo.favaro@unibe.ch using the following template:

   Subject: Application for Bern Face Dataset
   
   Content Template:
   Applicant: ________________________________________
   Occupation: _______________________________________
   Institution: ______________________________________
   Purpose of Usage: _________________________________
   
   Attachment: Scan of signed license agreement
         

Recommendations

• Apply using a *.edu e-mail address, which is more likely to be authorized.
• Attach a brief resume or personal page URL to help us identify you. If we cannot identify you as a professor or researcher, the application may be ignored.

Note

We find that many of our replies are mislabeled as spam, so if you don't get a response more than 72 hours after sending the request on a working day, please check the spam box just in case.