• Home
  • Search
  • TML: A Triple-Wise Multi-Task Learning Framework for Distracted Driver Recognition
  • Cite Icon12
  • https://doi.org/10.1109/access.2021.3109815Copy DOI Icon

TML: A Triple-Wise Multi-Task Learning Framework for Distracted Driver Recognition

Show More
  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

We propose a multi-task learning framework for improving the performance of vision-based deep-learning approaches for driver distraction recognition. The most popular tool so far for solving this task is convolutional neural networks (CNNs) that have proven to be strongly biased toward local features. Such bias causes CNNs to neglect global structural information, adversely affecting the robustness of the distracted driver recognition task. To solve this problem, we generate positive and negative samples of each given input, and construct a triplet of images (i.e., raw image, positive sample, and negative sample). The positive sample is generated by applying structure-aware illumination to the human body region of each given input. The negative sample is generated by randomly shuffling the local regions of each given input. The networks are then trained with the triplets using a multi-task learning strategy to force the networks to explore global information by multiple tasks: (a) recognizing the raw input and positive sample as the given ground truth; (b) recognizing the negative sample as an extra “meaningless” label; (c) pulling closer the distance between the features obtained from the raw input and positive sample while pushing away the distance between the features obtained from the raw input and negative sample. By doing so, the model can be trained so that it neglects the background information and pays more attention to the global structual information of the scene. The proposed approach reaches state-of-the-art performance on the AUC Distracted Driver Dataset and performs better than state-of-the-art studies on the Drive and Act Dataset. With raw images as input, we have achieved an accuracy of 96.0% for the AUC distracted driver dataset and 66.8% for the Drive and Act Dataset. Our approach does not introduce extra overhead during the testing procedure (i.e., utilization procedure), which is helpful for real-life applications. Moreover, better accuracy can be achieved by fusing the predictions respectively obtained with the raw input and positive sample. As a result, we have achieved an accuracy of 96.3% for the AUC distracted driver Dataset and 66.9% for the Drive and Act Dataset. The class activation map (CAM) of our proposed method is subjectively more reasonable, which would enhance the reliability and explainability of the model.

Similar Papers
  • Conference Article
  • Citations31

Dynamic cascades with bidirectional bootstrapping for spontaneous facial action unit detection

  • Sep 01, 2009
  • Yunfeng Zhu +3
  • Research Article
  • Citations22

Development and Comparison of a Rapid Isothermal Nucleic Acid Amplification Test for Typing of Herpes Simplex Virus Types 1 and 2 on a Portable Fluorescence Detector

  • Aug 27, 2012
  • The Journal of molecular diagnostics : JMD
  • Yanhong Tong +3
  • Research Article
  • Citations8

Real-time prediction of upcoming respiratory events via machine learning using snoring sound signal.

  • Apr 12, 2021
  • Journal of clinical sleep medicine : JCSM : official publication of the American Academy of Sleep Medicine
  • Bochun Wang +6
  • Research Article
  • Citations4

Feasibility of Hepatitis B virus infection detection using Mueller matrix decomposition method and Mueller matrix imaging polarimetry technique

  • Mar 12, 2022
  • Optics Communications
  • Le-Y Nguyen +6
  • PDF
  • Research Article
  • Citations4

Predicting potential miRNA-disease associations based on more reliable negative sample selection

  • Oct 17, 2022
  • BMC Bioinformatics
  • Ruiyu Guo +4
  • Research Article

Dual-track collaboration: Joint processing of heterogeneous positive and negative graph convolutional network for whole-slide image analysis.

  • Oct 01, 2025
  • Computer methods and programs in biomedicine
  • Meiyan Liang +7
  • Research Article
  • Citations11

PaaRPN: Probabilistic anchor assignment with region proposal network for visual tracking

  • Mar 23, 2022
  • Information Sciences
  • Kai Yang +3
  • Research Article
  • Citations27

Influence of Mastitis on Properties of Milk. V. Total and Free Concentrations of Major Minerals in Skimmilk

  • Oct 01, 1970
  • Journal of Dairy Science
  • P.T Tallamy +1
  • PDF
  • Research Article
  • Citations3

Multiple sampling schemes and deep learning improve active learning performance in drug-drug interaction information retrieval analysis from the literature

  • May 30, 2023
  • Journal of Biomedical Semantics
  • Weixin Xie +3
  • Research Article

Contrastive text embeddings with effective sample mining for enhanced disease diagnosis

  • Dec 01, 2025
  • Array
  • Xuanyi Zhang +6
  • Research Article

Deep one-class probability learning for end-to-end image classification.

  • May 01, 2025
  • Neural networks : the official journal of the International Neural Network Society
  • Jia Liu +4
  • Research Article

Do Antibacterial Milk Constituents Interfere with Microbiological Examination of Clinical Mastitis?

  • Jan 01, 2003
  • Acta Veterinaria Scandinavica
  • B.-A. Tenhagen +4
  • Research Article
  • Citations6

Leukocyte classification using relative-relationship-guided contrastive learning

  • Sep 13, 2024
  • Expert Systems With Applications
  • Zuoyong Li +5
  • Research Article
  • Citations9

Identification of Protein-Protein Interaction Associated Functions Based on Gene Ontology.

  • Mar 04, 2024
  • The protein journal
  • Yu-Hang Zhang +7
  • Research Article

Memory-guided mask reconstruction with central contrastive learning for robust multivariate time series anomaly detection.

  • Jul 01, 2026
  • Neural networks : the official journal of the International Neural Network Society
  • Le He +7
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.