and gesture information of a cartoon character. However
HEF feature is used to extract the shape of the cartoon
character, it is not efficient. Because of, affine
transformation. So we proposed CSS feature for
extracting shape with affine transformation. To
efficiently and effectively combine these visual features
for subsequent retrieval and synthesis tasks, we developed
a new multiview learning method named semi-MSL. It
encodes different features from different views in a
physically meaningful subspace. In this unified subspace,
the Euclidean distance can be directly used to measure the
distance between two cartoon characters. After ranking
similar cartoon characters can be retrieved. Experimental
evaluations on cartoon character retrieval and clip
synthesis suggest the effectiveness of the visual features
and semi-MSL.