We provide a qualitative analysis of the descriptions containing negations(no, not, n't, nobody, etc) in the Flickr30K corpus, and a categorization ofnegation uses. Based on this analysis, we provide a set of requirements that animage description system should have in order to generate negation sentences.As a pilot experiment, we used our categorization to manually annotatesentences containing negations in the Flickr30K corpus, with an agreement scoreof K=0.67. With this paper, we hope to open up a broader discussion ofsubjective language in image descriptions.
展开▼
