关键词: 汉明嵌入, 弱几何一致性约束, 词袋模型, 不确定性, 云模型

Abstract:

Because traditional duplicate image detection technologies can not ensure the scalability and the precision of image retrieval, this article proposes a duplicate image detection approach based on two-dimensional cloud model filter. On basis of bag-of-word model, the approach first maps the matching descriptors which are refined by Hamming embedding to points in the two-dimensional space, and then uses cloud model to compute the uncertainty of two-dimensional points' distribution for excluding the candidate images with larger fluctuation. Finally, the images are ranked according to voting score. Experiments show that the new approach not only maintains the merit of weak geometric consistency constraint algorithm which is suitable for large-scale image retrieval, but also significantly improves the accuracy of duplicate image detection.

Key words: Hamming embedding, weak geometric consistency constraint, bag-of-word, uncertainty, cloud model