Analyzing sedentary behavior in life-logging images
Research output: Contribution to journal › Conference article › Research › peer-review
Standard
Analyzing sedentary behavior in life-logging images. / Moghimi, Mohammad; Wu, Wanmin; Chen, Jacqueline; Godbole, Suneeta; Marshall, Simon; Kerr, Jacqueline; Belongie, Serge.
In: 2014 IEEE International Conference on Image Processing, ICIP 2014, 28.01.2014, p. 1011-1015.Research output: Contribution to journal › Conference article › Research › peer-review
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - GEN
T1 - Analyzing sedentary behavior in life-logging images
AU - Moghimi, Mohammad
AU - Wu, Wanmin
AU - Chen, Jacqueline
AU - Godbole, Suneeta
AU - Marshall, Simon
AU - Kerr, Jacqueline
AU - Belongie, Serge
N1 - Publisher Copyright: © 2014 IEEE.
PY - 2014/1/28
Y1 - 2014/1/28
N2 - We describe a study that aims to understand physical activity and sedentary behavior in free-living settings. We employed a wearable camera to record 3 to 5 days of imaging data with 40 participants, resulting in over 360,000 images. These images were then fully annotated by experienced staff with a rigorous coding protocol. We designed a deep learning based classifier in which we adapted a model that was originally trained for ImageNet [1]. We then added a spatio-temporal pyramid to our deep learning based classifier. Our results show our proposed method performs better than the state-of-the-art visual classification methods on our dataset. For most of the labels our system achieves more than 90% average accuracy across different individuals for frequent labels and more than 80% average accuracy for rare labels.
AB - We describe a study that aims to understand physical activity and sedentary behavior in free-living settings. We employed a wearable camera to record 3 to 5 days of imaging data with 40 participants, resulting in over 360,000 images. These images were then fully annotated by experienced staff with a rigorous coding protocol. We designed a deep learning based classifier in which we adapted a model that was originally trained for ImageNet [1]. We then added a spatio-temporal pyramid to our deep learning based classifier. Our results show our proposed method performs better than the state-of-the-art visual classification methods on our dataset. For most of the labels our system achieves more than 90% average accuracy across different individuals for frequent labels and more than 80% average accuracy for rare labels.
KW - Deep Learning
KW - Large Scale Image Analysis
KW - Visual Classification
KW - Wearable camera
UR - http://www.scopus.com/inward/record.url?scp=84949929440&partnerID=8YFLogxK
U2 - 10.1109/ICIP.2014.7025202
DO - 10.1109/ICIP.2014.7025202
M3 - Conference article
AN - SCOPUS:84949929440
SP - 1011
EP - 1015
JO - 2014 IEEE International Conference on Image Processing, ICIP 2014
JF - 2014 IEEE International Conference on Image Processing, ICIP 2014
ER -
ID: 302043908