Our goal is to develop the fundamental knowledge to design a visual system that is able to learn, recognize and retrieve quickly and accurately thousands of visual categories, including materials, objects, scenes, human actions and activities. A ``visual google'' for images and videos -- able to search for the ``nouns'' (objects, scenes), ``verbs'' (actions/activities) and adjectives (materials, patterns) of visual content. The time is right for making great progress in automated visual recognition: imaging geometry is well understood, image features are now highly developed, and relevant statistical models and machine learning algorithms are well-advanced. Our goal is to make a quantum leap in the capabilities of visual recognition in real-life scenarios. The outcomes of this research will impact any applications where visual recognition is useful, and will enable new applications entirely: effortlessly searching and annotating home image and video collections on their visual content; searching and annotating large commercial image and video archives (e.g. YouTube); surveillance; using an image, rather than text, to access the web and hence identify its visual content.