Multi-Task Learning for Fatigue Detection and Face Recognition of Drivers via Tree-Style Space-Channel Attention Fusion Network

TL;DR

This work tackles fatigue detection and driver identity recognition from facial images in driving scenarios, addressing redundancy in parallel multi-task models. It proposes a tree-style multi-task architecture (T-SCAF) with a shared backbone and two LASE-Net branches that jointly learn fatigue detection and face recognition by fusing space and channel attention. To train on existing single-task datasets, it introduces alternating updation and gradient accumulation, optimizing $L = wL_{Drowsy} + (1-w)L_{Face}$ with $w$ in [0,1], and losses including cross-entropy for drowsiness and ArcFace for identity. Experiments on a self-built fatigue dataset, CASIA-WebFace, and LFW show improved fatigue detection accuracy and competitive face recognition with fewer parameters and computation than parallel-style baselines. The approach highlights the value of tree-style sharing and attention fusion for efficient, accurate driver safety systems.

Abstract

In driving scenarios, automobile active safety systems are increasingly incorporating deep learning technology. These systems typically need to handle multiple tasks simultaneously, such as detecting fatigue driving and recognizing the driver's identity. However, the traditional parallel-style approach of combining multiple single-task models tends to waste resources when dealing with similar tasks. Therefore, we propose a novel tree-style multi-task modeling approach for multi-task learning, which rooted at a shared backbone, more dedicated separate module branches are appended as the model pipeline goes deeper. Following the tree-style approach, we propose a multi-task learning model for simultaneously performing driver fatigue detection and face recognition for identifying a driver. This model shares a common feature extraction backbone module, with further separated feature extraction and classification module branches. The dedicated branches exploit and combine spatial and channel attention mechanisms to generate space-channel fused-attention enhanced features, leading to improved detection performance. As only single-task datasets are available, we introduce techniques including alternating updation and gradient accumulation for training our multi-task model using only the single-task datasets. The effectiveness of our tree-style multi-task learning model is verified through extensive validations.

Figures (6)

• Figure 1: Structure of our tree-style space-channel attention fusion (T-SCAF) model for fatigue detection and face recognition of drivers.
• Figure 2: Images in drowsy and no-drowsy status in the self-built dataset
• Figure 3: Confusion Matrix
• Figure 4: Visualization of LASE-Net output feature maps
• Figure 5: t-SNE visualization results.