Fine-grained action segmentation using the semi-supervised action GAN

作者：

Highlights：

• A novel deep neural memory architecture for sequence to sequence modelling.

• Capture inter (long term) and intra (short term) sequence relationships.

• Thorough evaluation on air traffic modelling and pedestrian trajectory prediction.

• Qualitative evaluations show how tree memory structure captures dependencies.

摘要

•A novel deep neural memory architecture for sequence to sequence modelling.•Capture inter (long term) and intra (short term) sequence relationships.•Thorough evaluation on air traffic modelling and pedestrian trajectory prediction.•Qualitative evaluations show how tree memory structure captures dependencies.

论文关键词：Human action segmentation,Generative adversarial networks,Context modelling

论文评审过程：Received 27 October 2018, Revised 3 September 2019, Accepted 4 September 2019, Available online 5 September 2019, Version of Record 12 September 2019.

论文官网地址：https://doi.org/10.1016/j.patcog.2019.107039