Computer vision is one of the leading areas of AI, and has made significant progress in recent years. Computers can now generate images from text prompts, detect objects in images, isolate audio and visual streams in video, and match or classify the content of images of all kinds. Increasingly, computer vision is an element of multimodal learning, in which information belonging to multiple human senses is combined. This presentation and workshop will outline the state of the art in the field, with particular reference to humanities applications. Participants will have the opportunity to try hands-on demos based on current research collaborations with Oxford’s Visual Geometry Group; to learn for themselves what can be done with current technology; and to discuss such critical and ethical issues as privacy, bias and the limits and capabilities of machine understanding as a whole.
Methods Workshop Series requirements: This workshop is aimed at beginners. It is useful if participants have basic computer skills, but nothing more is required! If you wish to participate, do please bring along your device (laptop/tablet).
