Hierarchical Annotation Management Method for Activity Information Collecting System

Yuichi Hattori, Shota Tanaka, Sozo Inoue,
International Workshop on Network Traffic Control, Analysis and Applications (NTCAA-2012)
(Not Available)
(Not Available)
Victoria, Canada
Recently, researches for activity recognition with sensor data have been increased by the spread of smart phones with 3-axis accelerometers and gyroscopes. However, collecting various sensor data from many people is difficult. To solve this problem, we developed activity collecting platform using various client software. Moreover, annotating continuous activity data is also an important task. However, this task requires a lot of work by hands. Moreover, an activity has a complex structure, and annotation by human often contains ambiguity. For this reason, a system needs to support work so that a worker may add annotation efficiently and accuracy. Therefore, we propose the annotation software which has the function of hierarchical segmentation and hierarchical annotation.

Data Files