Dual-stream multidimensional network for aluminum electrolysis process superheat identification
Authors:
- 25 November 2022
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2873-2878
Abstract
Superheat identification is very important in aluminum electrolysis industry. In order to realize the real-time identification of superheat, this paper proposes a novel dual-stream multidimensional network (DM) model, which is composed of an efficient stream and an accuracy stream. In efficient stream, a mobilenet is equipped on Artificial Intelligence of Things (AIoT) devices to achieve quick superheat identification as normality or abnormality. In accuracy stream, a two-stream information sharing 3D residual network is equipped on a cloud computing service to achieve accurate identification through abnormal frames. Experiments are conducted on the aluminum electrolysis datasets, and the results verify that our DM model can realize efficient and accurate identification under resource saving condition.Keywords
Funding Information
- Nature
- National Natural Science Foundation of China
This publication has 9 references indexed in Scilit:
- Multi-cue based four-stream 3D ResNets for video-based action recognitionInformation Sciences, 2021
- Motion saliency based multi-stream multiplier ResNets for action recognitionImage and Vision Computing, 2021
- Information synergy entropy based multi-feature information fusion for the operating condition identification in aluminium electrolysisInformation Sciences, 2021
- One-Shot Learning for Surveillance Anomaly Recognition using Siamese 3D CNN2020
- Multi‐cue combination network for action‐based video classificationIET Computer Vision, 2019
- Building edge intelligence for online activity recognition in service-oriented IoT systemsFuture Generation Computer Systems, 2018
- Spatiotemporal Multiplier Networks for Video Action Recognition2017
- Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset2017
- Learning Spatiotemporal Features with 3D Convolutional Networks2015