Semi‐supervised long short‐term memory for human action recognition. Issue 13 (27th July 2020)

• Record Type: Journal Article
• Authors: Liu, Hong; Liu, Chang; Ding, Runwei
• Is Part Of: Journal of engineering. Volume 2020:Issue 13(2020)
• Page Start: 373
• Page End: 378
• Publication Date: 2020-07-27
• Subjects: learning (artificial intelligence) -- image recognition -- image motion analysis -- recurrent neural nets; human action recognition task -- unlabelled samples -- labelled samples -- semisupervised learning research -- semisupervised methods -- LSTM ladder network -- learning performance -- semisupervised long short‐term memory -- symmetrical LSTM network -- discriminative hidden features; Engineering -- Periodicals; Engineering; Electronic journals; Periodicals; 620.005
• DOI: 10.1049/joe.2019.1166
• Languages: English
• ISSNs: 2051-3305

Abstract

Abstract : In real human action recognition task, it is a common phenomenon that there are many unlabelled samples and few labelled samples. How to make good use of unlabelled samples to improve the generalisation ability of models is the focus of semi‐supervised learning research. In this study, the authors present two semi‐supervised methods based on long short‐term memory (LSTM) to learn discriminative hidden features. One is the LSTM ladder network, the other is the Symmetrical LSTM network. By them unlabelled samples can be used automatically to improve learning performance without relying on external interaction. Both on the NTU‐RGB + D dataset and the Kinetics dataset, their methods achieve >10 and 5% improvements, separately.