Why does the shower curtain always stick to your legs? What is more sustainable, cheese or herring? These kind of questions were asked in the annual national science quiz (de Nationale Wetenschapsquiz 2017), which aired the 26th of December (watch it here). Ionica Smeets and Pieter Hulst presented the quiz and the guests were three duos consisting of a scientist and an artist. In this program, the teams tried to answer surprising and interesting questions on scientific research. You can find all the questions and answers here (in Dutch). Two teams played the finale, where the program used our FaceReader software in a very interesting way.

In many futuristic movies, you see robots performing countless day-to-day tasks. Well… the future is here (almost)! For a project funded by COMMIT, we helped create a robot receptionist, named R3D3 (Rolling Receptionist Robot with Double Dutch Dialogue). The aim of this project was to create a combination of a virtual human and a robot capable of verbal and non-verbal interactions with humans. Together with University of Twente’s HMI and RAM, we succeeded in building a robot platform with the technical capacities to realize such interactions.

The R3D3 prototype can drive around, adjust its height, and carries a tablet with a virtual human face. The robot includes technology for speech recognition and speech production, and has FaceReader based computer vision techniques that can recognize gender, age and emotions. In addition, the virtual avatar on the tablet can interact with people. Here we report the results of three pilot studies, carried out to evaluate the performance of the robot and investigate how people reacted to it. Each pilot tested a different target population; shop visitors, police personnel, and children.

What does lust look like? How do people respond to something they find exciting? FaceReader has endless application possibilities. The Dutch television program “Spuiten en Slikken” used FaceReader to measure responses towards explicit erotic images. Spuiten en Slikken is a spicy television show, on a public service channel, that informs their viewers about everything related to sex and drugs. In this current topic, which aired on Tuesday the 24th (warning: explicit content), they interviewed a girl who was asexual. Someone who is asexual lacks sexual attraction to others. Sexologists consider it a sexual orientation, but there is a lot of prejudice towards it. Therefore, the TV show used different ways to figure out what it means to be asexual. One of these was using FaceReader to measure someone’s objective response towards erotic videos.

A few weeks ago, VicarVision crossed the Atlantic and got to show its first validation results for Baby FaceReader at the Pangborn Sensory Science Symposium.

Have you ever wondered what your baby is thinking or feeling? Baby FaceReader automatically measures facial expressions in infants (0-2 years old) to determine just that! The widely used FaceReader software can currently only be used for children from the age of three. VicarVision has been developing Baby FaceReader as part of the Brainview Marie Sklodowska-Curie European Training Network. We are proud to share its first promising validation results!

^{Example of a child where a few activated action units are scored.}

Last week the Imperial College of London hosted the 28th edition of the British Machine Vision Conference (BMVC). From all over the world researchers and representatives of industry (Google, Microsoft, Facebook, etc.) met up to present and discuss the newest developments in the fields of Computer Vision and Machine Learning. Since only 188 of the 635 submitted papers were accepted to this event, we were proud to attend and present some of the research done at VicarVision.

^{We were represented by  Nicolai and Amogh, who presented their paper “Object Extent Pooling for Weakly Supervised Single-Shot Localization” at BMVC’17.}

In products like FaceReader and Vicar Analytics we make use of the modern algorithms that are usually referred to as “Deep Learning”. These algorithms have shown to work extremely well for finding objects (i.e. faces, people) in images or video. The work that we presented at BMVC shares one of our new ideas about how we can make these algorithms much faster and lighter (in terms of computer memory) while not having to trade against accuracy. At the conference this idea was well received and we had the opportunity for a lot of interesting discussion, which will contribute to further development of our algorithms.

Aside from inspiring research, BMVC’17 also provided a podium to some interesting developments in the industry. Richard Szeliski for instance, Director of the Computational Photography Group at Facebook, demonstrated what his team is currently working on.