Technical Update – Better Analysis Results with FaceReader 9

We have only recently released the new desktop version FaceReader 9, but we did not want to keep those amazing improvements from the FaceReader Online user. So, we have already updated the FaceReader Online engine to FaceReader 9.

Improvements for online data collection

When people are at home they often have less ideal recording surroundings compared to the lab environment (see this blog for online/offline trade-offs). Sometimes people move away from the camera or do not have optimal lighting conditions. We have added a new deep learning based face modelling technique that is now able to handle bad lighting and different poses much better. We have also retrained our deep neural networks and increased the size of our training sets, which creates more robust facial expressions classification. When you start a new project FaceReader 9 is now selected as a default (but you can also select FaceReader 8, if you prefer, e.g. for comparison purposes).

Low quality results do not cost you credits

On top of this, we have decided that recordings that receive a quality score below 3, are now not deducted from your credits (this is now automatically incorporated). So, you could use those credits to test new participants. Of course, it is still important to think about how you can get good quality data (see this blog on how to optimize this). But we want to do everything in our power to help you get good results.