TY - GEN
T1 - Temporal alignment improves feature quality
T2 - 31st Meeting of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2018
AU - Choi, Hongjun
AU - Wang, Qiao
AU - Toledo, Meynard
AU - Turaga, Pavan
AU - Buman, Matthew
AU - Srivastava, Anuj
N1 - Funding Information:
The work of HC, QW, PT was supported by NSF grants 1452163and1617999. TheworkofASwassupportedby NSF grant 1617397.
Publisher Copyright:
© 2018 IEEE.
PY - 2018/12/13
Y1 - 2018/12/13
N2 - Activity recognition has been receiving significant attention from a variety of research areas such as human performance enhancement, health promotion, and human computer interaction. However, recognizing activities from accelerometer data still remains a challenging problem due to sensitivity to sampling rates, misalignment of data, and increased variability in activities among clinically relevant populations. In order to solve these issues, we adopt methods from functional analysis, which consider non-elastic rate variations in movement. The overall framework factors out temporal variability within activity classes, before leveraging robust machine learning pipelines for a given end-use. The proposed approach has been evaluated on 7 classes of everyday activities with 50 subjects. The results indicate that proposed approach achieves improved performance with the improvements observed in separating similar classes that differ in temporal rates, and also demonstrate higher robustness to change in window lengths. These results suggest that temporal alignment should be considered a core part of activity recognition pipelines.
AB - Activity recognition has been receiving significant attention from a variety of research areas such as human performance enhancement, health promotion, and human computer interaction. However, recognizing activities from accelerometer data still remains a challenging problem due to sensitivity to sampling rates, misalignment of data, and increased variability in activities among clinically relevant populations. In order to solve these issues, we adopt methods from functional analysis, which consider non-elastic rate variations in movement. The overall framework factors out temporal variability within activity classes, before leveraging robust machine learning pipelines for a given end-use. The proposed approach has been evaluated on 7 classes of everyday activities with 50 subjects. The results indicate that proposed approach achieves improved performance with the improvements observed in separating similar classes that differ in temporal rates, and also demonstrate higher robustness to change in window lengths. These results suggest that temporal alignment should be considered a core part of activity recognition pipelines.
UR - http://www.scopus.com/inward/record.url?scp=85060897066&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85060897066&partnerID=8YFLogxK
U2 - 10.1109/CVPRW.2018.00075
DO - 10.1109/CVPRW.2018.00075
M3 - Conference contribution
AN - SCOPUS:85060897066
T3 - IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops
SP - 462
EP - 470
BT - Proceedings - 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2018
PB - IEEE Computer Society
Y2 - 18 June 2018 through 22 June 2018
ER -