Monday, 2021/01/25

  • CS/AI
  • C
  • TITECH
  • Switch Language
    • ja日本語 (Japanese)
    • enEnglish

Shimosaka Research Group

pursuing MIUBIQ (machine intelligence in UbiComp Research)

  • Home
    • Members
    • Location
  • News
  • Projects
  • Publications
  • Awards
  • Archives
    • Codes
    • Datasets
Navigation
News Presenting our paper on neural-based time-series multi-label classification problem towards reliable activity recognition on construction vehicles at IROS2020

Presenting our paper on neural-based time-series multi-label classification problem towards reliable activity recognition on construction vehicles at IROS2020

2020/10/26 | NewsPresentations | 164 views |

The 2020 IEEE/RSJ International Conference on Intelligent Robots (IROS2020) is now held October 25 – November 25 at on-demand conference. The following oral presentation will be delivered.

Sensor-based activity recognition for construction vehicles is useful for evaluating the skills of the operator, measuring work efficiency, and many other use cases. Therefore, many researches have explored robust activity-recognition models. However, it remains a challenge to apply the model to many construction sites because of the imbalance of the dataset.
While it is natural to employ multi-label representation on imbalanced data with a large number of activity categories, multi-label robust classification for activity recognition has yet to be resolved because of the nature of the time-series property.

In this work, we propose a novel multi-label long short-term memory (LSTM) model, which is effective for the sequence multi-labeling problem. The proposed model has connections to the temporal direction and attribute direction, which exploit both the temporal pattern and co-occurrence among attributes. In addition, by providing a bidirectional connection structure in the attribute direction, the model enables us to alleviate the dependency of the chain order in what we call classifier chain, which is a classical approach to multi-label classification. To validate our methods, we conduct experiments using real-world construction-vehicle dataset.

-presentation information-
Haruka Abe*, Takuya Hino**, Motohide Sugihara**, Hiroki Ikeya**, and Shimosaka, Masamichi*: Multi-Label Long Short-Term Memory for Construction Vehicle Activity Recognition with Imbalanced Supervision.
(*Tokyo Institute of Technology, **Komatsu Ltd.)

NOTE: You could register IROS2020 on-demand for free via

  • tweet

Comments are disabled for this post

Social Networks

  • twitter
  • rss

Recent News

  • Presenting our paper on fine-grained urban anomaly detection at nationwide scale using massive GPS logs on BigData2020 2020/12/11
  • GroupWiLo: Accurate and Low-cost Wi-Fi-based Localization with Minimization of Parameter Total Variation 2020/11/10
  • Adaptive driving behavior modeling towards diverse driving situations using inverse reinforcement learning 2020/11/02
  • Presenting our paper on accurate anomaly crowd prediction at ACM SIGSPATIAL 2020 2020/10/27
  • Presenting our paper on neural-based time-series multi-label classification problem towards reliable activity recognition on construction vehicles at IROS2020 2020/10/26
  • Presenting our paper on Wi-Fi-based indoor localization maintenance method via group-wise total variation regularization at IPIN2019 2019/10/01
  • Presenting our project on VR art “CityAtmosphere” in UbiComp 2019 Demos. 2019/09/04
  • Presenting our paper on large-scale fine-grained urban dynamics prediction at AAAI2019 2019/01/28
  • Presenting our paper on population pattern prediction at ACM SIGSPATIAL2018 2018/11/01
  • New members of September 2018 2018/10/19

Search

Copyright 2015 · Shimosaka Research Group at TITECH