Abstract: Most of the existing approaches for image annotation generally demand exactly labeled training data, which are often difficult to obtain. In this letter we present a novel model that ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results