Deep Convolutional Bidirectional LSTM for Complex Activity Recognition with Missing Data

Swapnil Sayan Saha; Sandeep Singh Sandha; Mani Srivastava

doi:10.1007/978-981-15-8269-1_4

Deep Convolutional Bidirectional LSTM for Complex Activity Recognition with Missing Data

Swapnil Sayan Saha, Sandeep Singh Sandha, Mani Srivastava

Research output: Chapter in Book/Report/Conference proceeding › Chapter

8 Scopus citations

Abstract

Complex activity recognition using multiple on-body sensors is challenging due to missing samples, misaligned data times tamps across sensors, and variations in sampling rates. In this paper, we introduce a robust training pipeline that handles sampling rate variability, missing data, and misaligned data time stamps using intelligent data augmentation techniques. Specifically, we use controlled jitter in window length and add artificial misalignments in data timestamps between sensors, along with masking representations of missing data. We evaluate our pipeline on the Cooking Activity Dataset with Macro and Micro Activities, benchmarking the performance of deep convolutional bidirectional long short-term memory (DCBL) classifier. In our evaluations, DCBL achieves test accuracies of 88% and 72%, respectively, for macro- and micro-activity classifications, exceeding performance over state-of-the-art vanilla activity classifiers.

Original language	English (US)
Title of host publication	Human Activity Recognition Challenge
Publisher	Springer Singapore
Pages	39-53
Number of pages	15
ISBN (Print)	9789811582684
DOIs	https://doi.org/10.1007/978-981-15-8269-1_4
State	Published - Nov 21 2020
Externally published	Yes

Bibliographical note

KAUST Repository Item: Exported on 2021-07-01
Acknowledged KAUST grant number(s): Sensor Innovation research program
Acknowledgements: The research reported in this paper was sponsored in part by the CONIX Research Center, one of six centers in JUMP, a Semiconductor Research Corporation (SRC) program sponsored by DARPA, the Army Research Laboratory (ARL) under Cooperative Agreement W911NF-17-2-0196 and the King Abdullah University of Science and Technology (KAUST) through its Sensor Innovation research program. Any findings in this material are those of the author(s) and do not reflect the views of any of the above funding agencies. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation hereon.
This publication acknowledges KAUST support, but has no KAUST affiliated authors.

Access to Document

10.1007/978-981-15-8269-1_4

Cite this

@inbook{b7fb85af56ef44a281153593d2ad0a56,

title = "Deep Convolutional Bidirectional LSTM for Complex Activity Recognition with Missing Data",

abstract = "Complex activity recognition using multiple on-body sensors is challenging due to missing samples, misaligned data times tamps across sensors, and variations in sampling rates. In this paper, we introduce a robust training pipeline that handles sampling rate variability, missing data, and misaligned data time stamps using intelligent data augmentation techniques. Specifically, we use controlled jitter in window length and add artificial misalignments in data timestamps between sensors, along with masking representations of missing data. We evaluate our pipeline on the Cooking Activity Dataset with Macro and Micro Activities, benchmarking the performance of deep convolutional bidirectional long short-term memory (DCBL) classifier. In our evaluations, DCBL achieves test accuracies of 88% and 72%, respectively, for macro- and micro-activity classifications, exceeding performance over state-of-the-art vanilla activity classifiers.",

author = "Saha, {Swapnil Sayan} and Sandha, {Sandeep Singh} and Mani Srivastava",

note = "KAUST Repository Item: Exported on 2021-07-01 Acknowledged KAUST grant number(s): Sensor Innovation research program Acknowledgements: The research reported in this paper was sponsored in part by the CONIX Research Center, one of six centers in JUMP, a Semiconductor Research Corporation (SRC) program sponsored by DARPA, the Army Research Laboratory (ARL) under Cooperative Agreement W911NF-17-2-0196 and the King Abdullah University of Science and Technology (KAUST) through its Sensor Innovation research program. Any findings in this material are those of the author(s) and do not reflect the views of any of the above funding agencies. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation hereon. This publication acknowledges KAUST support, but has no KAUST affiliated authors.",

year = "2020",

month = nov,

day = "21",

doi = "10.1007/978-981-15-8269-1_4",

language = "English (US)",

isbn = "9789811582684",

pages = "39--53",

booktitle = "Human Activity Recognition Challenge",

publisher = "Springer Singapore",

address = "Singapore",

}

TY - CHAP

T1 - Deep Convolutional Bidirectional LSTM for Complex Activity Recognition with Missing Data

AU - Saha, Swapnil Sayan

AU - Sandha, Sandeep Singh

AU - Srivastava, Mani

N1 - KAUST Repository Item: Exported on 2021-07-01 Acknowledged KAUST grant number(s): Sensor Innovation research program Acknowledgements: The research reported in this paper was sponsored in part by the CONIX Research Center, one of six centers in JUMP, a Semiconductor Research Corporation (SRC) program sponsored by DARPA, the Army Research Laboratory (ARL) under Cooperative Agreement W911NF-17-2-0196 and the King Abdullah University of Science and Technology (KAUST) through its Sensor Innovation research program. Any findings in this material are those of the author(s) and do not reflect the views of any of the above funding agencies. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation hereon. This publication acknowledges KAUST support, but has no KAUST affiliated authors.

PY - 2020/11/21

Y1 - 2020/11/21

N2 - Complex activity recognition using multiple on-body sensors is challenging due to missing samples, misaligned data times tamps across sensors, and variations in sampling rates. In this paper, we introduce a robust training pipeline that handles sampling rate variability, missing data, and misaligned data time stamps using intelligent data augmentation techniques. Specifically, we use controlled jitter in window length and add artificial misalignments in data timestamps between sensors, along with masking representations of missing data. We evaluate our pipeline on the Cooking Activity Dataset with Macro and Micro Activities, benchmarking the performance of deep convolutional bidirectional long short-term memory (DCBL) classifier. In our evaluations, DCBL achieves test accuracies of 88% and 72%, respectively, for macro- and micro-activity classifications, exceeding performance over state-of-the-art vanilla activity classifiers.

AB - Complex activity recognition using multiple on-body sensors is challenging due to missing samples, misaligned data times tamps across sensors, and variations in sampling rates. In this paper, we introduce a robust training pipeline that handles sampling rate variability, missing data, and misaligned data time stamps using intelligent data augmentation techniques. Specifically, we use controlled jitter in window length and add artificial misalignments in data timestamps between sensors, along with masking representations of missing data. We evaluate our pipeline on the Cooking Activity Dataset with Macro and Micro Activities, benchmarking the performance of deep convolutional bidirectional long short-term memory (DCBL) classifier. In our evaluations, DCBL achieves test accuracies of 88% and 72%, respectively, for macro- and micro-activity classifications, exceeding performance over state-of-the-art vanilla activity classifiers.

UR - http://hdl.handle.net/10754/669846

UR - http://link.springer.com/10.1007/978-981-15-8269-1_4

UR - http://www.scopus.com/inward/record.url?scp=85096581056&partnerID=8YFLogxK

U2 - 10.1007/978-981-15-8269-1_4

DO - 10.1007/978-981-15-8269-1_4

M3 - Chapter

SN - 9789811582684

SP - 39

EP - 53

BT - Human Activity Recognition Challenge

PB - Springer Singapore

ER -

Deep Convolutional Bidirectional LSTM for Complex Activity Recognition with Missing Data

Abstract

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this