ZFZenan Fu

Zenan Fu

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Zenan Fu has a total of 15 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Rep-MMB: Bridging Mobile CNN and Transformer for Sensor-Based Human Activity Recognition

    by Jinsheng Liu, Lei Zhang, Xin Liu, Guangjie Chen, Zenan Fu, Wenbo Huang, Hao Wu, Aiguo Song


  • Open Access

    Domain Generalization for Sensor-Based Activity Recognition via Decomposed Feature Representation and Semantic Augmentation

    by Zenan Fu, Lei Zhang, Di Xiong, Wenbo Huang, Xin Liu

    Domain generalization (DG) has recently offered a promising solution to tackle distribution-shift problem in sensor-based human activity recognition (HAR) scenario. Particularly, the sensor data augmentation has gained a lot of attention, which aims to generate more diverse representations to learn model invariance across different domains. However, most existing methods seek to directly perform augmentation in the sensor input space, which result in a limited variety of augmented features. Intuitively, performing augmentation in feature space would be more versatile and diverse, which has been seldom considered in sensor-based cross-domain activity recognition. In this paper, we propose a new feature space augmentation method by decomposing the obtained features into four main components, i.e., category-invariant domain-invariant, category-invariant domain-specific, category-specific domain-invariant, and category-specific domain-specific, which enable us to increase sensor sample diversity while emphasizing invariance learning to enhance domain generalization for activity recognition. Extensive experiments on four public HAR benchmarks including DSADS, USC-HAD, PAMAP2 and UCI-HAR demonstrate that our proposed feature space augmentation approach is able to achieve state-of-the-art performance under Cross-person, Cross-dataset, Cross-position and One-person-to-another settings. Detailed ablation and visualizations analyses verify its effectiveness, efficiency, and universality under cross-domain activity recognition scenario.

  • TSA-Former: Linear Transformer with Taylor Series Attention for Sensor-Based Human Activity Recognition

    by Qifan Sun, Zechen Li, Kun Wang, Zenan Fu, Guangjie Chen, Lei Zhang, Hao Wu, Aiguo Song

    Transformer models have demonstrated superior capability in capturing long-range temporal dependencies crucial for Sensor-Based Human Activity Recognition (HAR). However, the quadratic computational complexity inherent to the Softmax-Attention mechanism significantly impedes their deployment on resource-constrained wearable devices and real-time streaming tasks. To address this, we propose a novel Linear Transformer with Taylor Series Attention specifically tailored for the HAR domain, named TSA-Former. It leverages the first-order Taylor expansion to approximate the Softmax-Attention and utilizes the norm-preserving mapping to approximate the high-order non-linear information, resulting in a linear computational complexity. In addition, TSA-Former integrates a multi-branch architecture featuring multi-scale patch embedding, which enables the model to dynamically capture multi-scale temporal features while minimizing overhead. Experimental results across four public HAR benchmarks, namely UniMiB-SHAR, UCI-HAR, WISDM, and OPPORTUNITY, demonstrate that TSA-Former achieves state-of-the-art (SOTA) accuracy and efficiency, outperforming conventional Transformers and existing linear-attention models. Deployment experiments conducted on the Raspberry Pi 5 platform further validate the model’s superior low-latency and minimal power consumption profile, confirming its robust suitability for real-world embedded HAR applications. Code will be released.


  • Cross-Modal Feature Adapter for Few-Shot Human Activity Recognition

    by Xin Liu, Lei Zhang, Wenbo Huang, Jinsheng Liu, Zenan Fu, Hao Wu, Aiguo Song

    Recent years have witnessed outstanding success of deep learning in sensor-based human activity recognition (HAR), spanning a wide range of real-world applications like healthcare management, fitness tracking, and fall detection. However, sensor data annotation scarcity still remains a main challenge unresolved, which requires human annotators to take a long-term and tedious observation to segment and timestamp sensor samples meticulously, hampering the wide use of deep learning models, especially in few-shot HAR scenario. To handle such issue, this paper introduces a cross-modal data augmentation, by exploiting activity label text as key words to search for activity-related images to construct an augmented dataset. On this basis, a new cross-modal feature adapter is designed, which fine-tunes a pre-trained CLIP image encoder to effectively align image-sensor pairs. Through a learnable residual ratio, it may adaptively blend the old knowledge inherited from the original zero-shot CLIP with the new knowledge adapted from few-shot training samples, which makes training converges faster whilst forming a streamlined time series sensor encoder. Extensive experiments and ablation studies are performed on three public HAR benchmarks. The experimental results demonstrate that the proposed method outperforms existing state-of-the-art HAR baselines under all few-shot scenarios. A practical on-device inference latency is provided.

  • ActiFormer: Sign-Aware Linear Attention for Sensor-Based Human Activity Recognition

    by Qifan Sun, Zenan Fu, Lei Zhang, Guangjie Chen, Wenbo Huang, Hao Wu, Aiguo Song

    Human Activity Recognition (HAR) plays a pivotal role in ubiquitous computing. However, it remains constrained by the challenge of balancing fine-grained temporal modeling with real-time efficiency on resource-limited devices. While Transformer-based models excel at capturing long-range dependencies, they suffer from high computational costs, limiting their applicability on resource-constrained devices. Linear attention mechanisms improve efficiency but often discard negative signals and produce overly smooth, high-entropy attention distributions, impairing the extraction of fine-grained patterns and degrading classification accuracy in complex scenarios. In this work, we present ActiFormer, a novel sign-aware linear attention framework tailored for sensor-based HAR to overcome these limitations. To preserve bidirectional signal dynamics, we introduce Sign-Aware Attention, which explicitly models both same-sign and cross-sign interactions between queries and keys, effectively retaining negative signals crucial for accurate recognition. Furthermore, we propose a learnable entropy-scaling function that compensates for the exponential scaling effect lost in linear attention, originally provided by softmax, solving the high-entropy attention weight issue by amplifying the importance of critical temporal points. Extensive experiments on four benchmark HAR datasets demonstrate that ActiFormer consistently outperforms CNNs, standard Transformers, and state-of-the-art linear attention models, both in accuracy and efficiency. Its lightweight design supports real-time inference on edge devices such as the Raspberry Pi 5, highlighting its practical deployability in real-world applications.


  • Optimizing Accuracy-Efficiency Trade-Offs of On-Device Activity Inference With Star Operation

    by Guangjie Chen, Zenan Fu, Yetong Sha, Di Xiong, Lei Zhang, Hao Wu, Aiguo Song

    Lightweight convolution-based neural networks (CNNs) are well suited for sensor-based human activity recognition (HAR) applications on resource-constrained edge devices with faster inference speed. However, the convolutional kernels are often limited to a small window range, which can only capture local details in time series sensor data, thus preventing further performance boost. Though Introducing self-attention into convolution can help to handle long-range dependence well, it might significantly slow down actual activity inference speed, due to high computational cost. In this paper, we introduce a new learning paradigm (star operation) and then present a lightweight Dual-Branch High-Order Interactions (DbHoi) block, which is computationally friendly for mobile HAR deployment. The proposed DbHoi block may implicitly transform raw sensor inputs into high-dimensional non-linear features, but actually operate in a low-dimensional feature space (analogs to the design principle of polynomial kernel tricks), without incurring extra computational overhead. Extensive experiments are conducted on three public HAR benchmarks including UCI-HAR, UniMiB-SHAR, and OPPORTUNITY, which demonstrate that our suggested DbHoi can consistently surpass various meticulously designed lightweight networks such as MobileNet, ShuffleNet, and GhostNet. Detailed ablation studies, visualizing representations, and on-device latency analyses further validate our insights with regards to the star operation, while underscoring its practical merit in real-world HAR deployment.


  • Deep Learning Anytime Prediction via Enforcing Runtime Monotonicity for Early-Exit Activity Recognition

    by Xin Liu, Lei Zhang, Wenbo Huang, Di Xiong, Zenan Fu, Hao Wu, Aiguo Song

    During the past decade, deep neural networks (DNNs) have been widely deployed to a variety of mobile devices for sensor-based human activity recognition (HAR), where their computational budgets are usually dynamic and restricted. Anytime algorithms are especially suitable for such dynamic HAR computing environments, which can return a prediction at an arbitrary time during computation, whose predictive quality gradually evolves as a function of computation time. Early-exit DNNs have recently gained a lot of attention in the context of HAR, which can make intermediate predictions across various exits throughout the whole network. However, most existing early-exit HAR works do not work well under anytime setting, since their predictive quality with runtime monotonicity per exit could not be guaranteed as computation time evolves. To mitigate this issue, this paper proposes a lightweight post-hoc transformation using a formulation of Product-of-Experts, which aims to endow deep models with runtime monotonicity, hence paving an essential step towards truly anytime predictive modeling for early-exit activity recognition. The comprehensive experiments and ablation studies on three publicly available HAR benchmark datasets verify that our anytime predictor can ensure such runtime monotonicity behavior while preserving competitive accuracy on average, with minimal implementation overhead. To the best of our knowledge, this paper is the first to impose runtime monotonicity on predictive quality for early-exit activity recognition. An actual evaluation is performed on a mobile device.

  • Learning Sensor Sample-Reweighting for Dynamic Early-Exit Activity Recognition Via Meta Learning

    by Zenan Fu, Lei Zhang, Wenbo Huang, Dongzhou Cheng, Hao Wu, Aiguo Song

    During recent years, dynamic early-exit has provided a promising paradigm to improve the computational efficiency of deep neural networks by constructing multiple classifiers to let easy samples exit at shallow layers while avoiding redundant computations at deep exits, which has been seldom explored in the context of latency-aware human activity recognition (HAR) deployed on wearable devices. Particularly, most existing early-exit strategies have always treated all activity samples equally at each exit during training, which ignore such dynamic early-exit behavior at test-time, causing a potential mismatch between training and test. Intuitively, easy activity samples that often exit earlier at test-time should place more emphasis on the training loss of shallow classifiers, while hard activity samples should contribute more to the training loss of deep classifiers. To bridge this gap, this paper introduces a sample-reweighting approach for efficient activity inference, which employs a weight-predicting network to reweight the training loss of different activity samples at every exit. From a perspective of meta learning, a new optimization objective function is designed to jointly optimize both weight-predicting network and backbone network. We perform extensive experiments on three popular HAR benchmarks including UCI-HAR, WISDM, and UniMiB-SHAR, which demonstrate that while incorporating such test-time early-exit behavior into conventional training pipeline, it can consistently improve the accuracy-efficiency trade-offs under budgeted batch classification and anytime prediction patterns. Moreover, our approach has a natural advantage in handing class-imbalance HAR problem. Detailed ablation studies, visualized illustrations, and real hardware deployment are provided to support our statement. Code will be available at https://github.com/FuZenan/LtSRW-for-EEAR .