Multi-label learning for concept-oriented labels of product image data

作者：

Highlights：

• A FCL network is proposed to exploit features correlations with gram matrices.

• The FCL network improves the performance of concept-oriented label prediction.

• The multi-label learning problem is firstly applied in product designing field.

• A product image dataset for multi-label learning is constructed and released.

摘要

•A FCL network is proposed to exploit features correlations with gram matrices.•The FCL network improves the performance of concept-oriented label prediction.•The multi-label learning problem is firstly applied in product designing field.•A product image dataset for multi-label learning is constructed and released.

论文关键词：Multi-label learning,Concept-oriented labels,Product image data,Feature correlation learning,Gram matrices

论文评审过程：Received 26 July 2019, Revised 20 October 2019, Accepted 22 October 2019, Available online 31 October 2019, Version of Record 21 January 2020.

论文官网地址：https://doi.org/10.1016/j.imavis.2019.10.007