Text to visual synthesis with appearance models | IEEE Conference Publication | IEEE Xplore