CVPR Daily - Thursday

DAILY T h u r s d a y C.Vision for Fashion, Art and Design 21 Hui would like to see fashion and the retail process become more conversational. “Fashion retrieval is usually image-to-image retrieval,” she explains. “You have a single item. You retrieve for a single item. That’s not always feasible when you go shopping. In the future, you can imagine you might interact with your smart assistant . You just want to talk. It would be really nice if you could have a dialogue that can help the computer to learn. That’s one of the popular areas in this space now. Negar and I are both working on research that combines language and vision in the same application. One of our speakers, Devi Parikh , is one of the leading figures in the language and vision space. She is now extending her work to computational creativity .” between the different disciplines, new topics appear. There are a number of interesting speakers this year, including Tamara Berg , but Hui and Negar refuse to give in to our persistent demands to choose a personal highlight. “You will enjoy all of them!” Negar insists, laughing. However, she does relent to say how excited she is that Hui herself will be speaking: “ Hui is not only my colleague but also a good friend. The part that made us really connect and learn from each other was the fact that we were both working on the intersection of language and computer vision. I’m really interestedtohearmoreabouther vision on continuing this intersection .” “Language is so important if we want to push computational creativity forward.”