Deep cross-view convolutional features for view-invariant action recognition

Research output: Book chapter/Published conference paperConference paperpeer-review

1 Citation (Scopus)


Convolutional neural network (CNN) based approaches have proved very effective for recognizing actions from a fixed viewpoint. However, these approaches are not generalized for recognizing actions captured from arbitrary viewpoint. In this paper, we present a deep multi-view framework for cross-view action recognition. We integrate spatiotemporal convolutional features from multiple views using deep multi-view representation learning. It helps to extract deep discriminative cross-view convolutional features for action recognition from any arbitrary viewpoint. To speed-up action detection and recognition, we then, train a feature based correlation filter for each action class. The proposed framework helps to recognize actions across different view-points with increased accuracy. An extensive experimentation to evaluate the underlying design on four publicly available datasets indicates that problem of view variations in in a single action class can be solved by learning discriminative information from multiple view.
Original languageEnglish
Title of host publication2018 IEEE International Conference on Image Processing, Applications and Systems (IPAS)
Number of pages5
ISBN (Electronic)9781728102474
ISBN (Print)9781728102481
Publication statusPublished - May 2019
Event2018 IEEE International Conference on Image Processing, Applications and Systems (IPAS) - Inria Sophia Antipolis – Méditerranée, Sophia Antipolis, France
Duration: 12 Dec 201814 Dec 2018 (proceedings) (Call for papers)


Conference2018 IEEE International Conference on Image Processing, Applications and Systems (IPAS)
CitySophia Antipolis
OtherThe international Image Processing Applications and Systems conference is intended for grouping from all over the world challenging researchers, innovators, academicians, and practitioners in image processing theory and tools, for following high level tutorials, sharing their achievements, exchanging their experiences and discussing future orientations. The conference will also offer an opportunity to make a bridge between image processing researchers and people working in other application fields such as medical doctors, radiotherapists or industrial parts.
The conference is devoted to image processing, computer vision algorithms and applications. Papers should report high quality research and describe original contributions.
Internet address


Dive into the research topics of 'Deep cross-view convolutional features for view-invariant action recognition'. Together they form a unique fingerprint.

Cite this