Abstract: Learning from image-text data has demonstrated recent success for many recognition tasks, yet is currently limited to visual features or individual visual concepts such as objects. In this ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results