Comparison of state-of-the-art deep learning APIs for image multi-label classification using semantic metrics

作者：

Highlights：

• Performance comparison of state-of-the-art image multi-label classification APIs.

• Propose semantic metrics to face the challenge of different APIs’ object-class sets.

• Show the merits of semantic metrics in comparing APIs trained on various datasets.

摘要

•Performance comparison of state-of-the-art image multi-label classification APIs.•Propose semantic metrics to face the challenge of different APIs’ object-class sets.•Show the merits of semantic metrics in comparing APIs trained on various datasets.

论文关键词：Image multi-label classification comparison,Semantic evaluation,Deep learning,Image understanding

论文评审过程：Received 16 October 2019, Revised 10 June 2020, Accepted 10 June 2020, Available online 2 July 2020, Version of Record 10 July 2020.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.113656