Diverse image annotation with missing labels

作者：

Highlights：

• A new task called diverse image annotation with missing labels (or DIAML) is proposed, that is more relevant to the real-world image annotation task.

• A new k-nearest neighbours based algorithm, called Per-Label kNN (or PL-kNN), is proposed to address the challenges involved in the DIAML task.

• Different aspects of the proposed approach are analyzed and compared with competing state-of-the-art methods.

• Extensive experiments demonstrate the efficacy of the proposed approach.

摘要

•A new task called diverse image annotation with missing labels (or DIAML) is proposed, that is more relevant to the real-world image annotation task.•A new k-nearest neighbours based algorithm, called Per-Label kNN (or PL-kNN), is proposed to address the challenges involved in the DIAML task.•Different aspects of the proposed approach are analyzed and compared with competing state-of-the-art methods.•Extensive experiments demonstrate the efficacy of the proposed approach.

论文关键词：Image annotation,Diverse labels,Missing labels,Nearest neighbour

论文评审过程：Received 23 February 2018, Revised 27 April 2019, Accepted 8 May 2019, Available online 8 May 2019, Version of Record 10 May 2019.

论文官网地址：https://doi.org/10.1016/j.patcog.2019.05.018