Computer Vision News - August 2022

56 Women in Computer Vision have this natural interaction. These agents would help humans in the future. Don't you sometimes look at a smile saying, “ Will my machine understand that? ” How do you train your software so that it understands, even if it doesn’t look like a full smile? I would ask you then, “ What do you mean if you, as a human, see this as a smile? ” How do you see it? I think the question would be why you, as a human, decide that this person is smiling. Toddlers learn very early to recognize smiles. But computer vision is getting better and better and actually can do that. I mean, I ask you how we humans do it because I think it's a bit similar. By having many samples or maybe a few samples from the same person, a smile can be detected. Some of these have been already discovered. The smile, for example, is not a very difficult task anymore. There are lots of cameras that can detect a smile now. could be a machine. It could be a personal assistant. It could be an app on the phone, right? It can be anything. All these kinds of models can be embedded in them, and the models are usually based on data. So what is the ground truth? Yes, our goal is to understand how we humans interact with each other, how we can understand each other better and respond in a natural way. That's what we want the machines to be able to do, to augment human intelligence and help us in our daily life. There are these assistive robots, for example, that can help provide care for the elderly. The models can also transfer the experience of an expert. For example, I'm interested in applications in mental health. I build data-driven models that are based on expert judgment related to psychological distress or depression and the aim is to make these models accessible as the first line of support. I would say that most of this is data-driven modeling, and the main aim is really to Marwa, we are not in Giza anymore!