Rui Liu

Ye Zhang, Qi Chen, Wenyou Huang et al.

Detection Transformer (DETR) offers an end-to-end solution for object detection by eliminating hand-crafted components like non-maximum suppression. However, DETR suffers from inefficient query competition where multiple queries converge to similar positions, leading to redundant computations. We present Route-DETR, which addresses these issues through adaptive pairwise routing in decoder self-attention layers. Our key insight is distinguishing between competing queries (targeting the same object) versus complementary queries (targeting different objects) using inter-query similarity, confidence scores, and geometry. We introduce dual routing mechanisms: suppressor routes that modulate attention between competing queries to reduce duplication, and delegator routes that encourage exploration of different regions. These are implemented via learnable low-rank attention biases enabling asymmetric query interactions. A dual-branch training strategy incorporates routing biases only during training while preserving standard attention for inference, ensuring no additional computational cost. Experiments on COCO and Cityscapes demonstrate consistent improvements across multiple DETR baselines, achieving +1.7% mAP gain over DINO on ResNet-50 and reaching 57.6% mAP on Swin-L, surpassing prior state-of-the-art models.

Rui Liu, Zikang Wang, Peng Gao et al.

Autonomous systems have advanced significantly, but challenges persist in accident-prone environments where robust decision-making is crucial. A single vehicle's limited sensor range and obstructed views increase the likelihood of accidents. Multi-vehicle connected systems and multi-modal approaches, leveraging RGB images and LiDAR point clouds, have emerged as promising solutions. However, existing methods often assume the availability of all data modalities and connected vehicles during both training and testing, which is impractical due to potential sensor failures or missing connected vehicles. To address these challenges, we introduce a novel framework MMCD (Multi-Modal Collaborative Decision-making) for connected autonomy. Our framework fuses multi-modal observations from ego and collaborative vehicles to enhance decision-making under challenging conditions. To ensure robust performance when certain data modalities are unavailable during testing, we propose an approach based on cross-modal knowledge distillation with a teacher-student model structure. The teacher model is trained with multiple data modalities, while the student model is designed to operate effectively with reduced modalities. In experiments on $\textit{connected autonomous driving with ground vehicles}$ and $\textit{aerial-ground vehicles collaboration}$, our method improves driving safety by up to ${\it 20.7}\%$, surpassing the best-existing baseline in detecting potential accidents and making safe driving decisions. More information can be found on our website https://ruiiu.github.io/mmcd.

Xi Fu, Rui Liu, Aung Aung Phyo Wai et al.

Decoding gait dynamics from EEG signals presents significant challenges due to the complex spatial dependencies of motor processes, the need for accurate temporal and spectral feature extraction, and the scarcity of high-quality gait EEG datasets. To address these issues, we propose EEG2GAIT, a novel hierarchical graph-based model that captures multi-level spatial embeddings of EEG channels using a Hierarchical Graph Convolutional Network (GCN) Pyramid. To further improve decoding accuracy, we introduce a Hybrid Temporal-Spectral Reward (HTSR) loss function, which combines time-domain, frequency-domain, and reward-based loss components. Moreover, we contribute a new Gait-EEG Dataset (GED), consisting of synchronized EEG and lower-limb joint angle data collected from 50 participants over two lab visits. Validation experiments on both the GED and the publicly available Mobile Brain-body imaging (MoBI) dataset demonstrate that EEG2GAIT outperforms state-of-the-art methods and achieves the best joint angle prediction. Ablation studies validate the contributions of the hierarchical GCN modules and HTSR Loss, while saliency maps reveal the significance of motor-related brain regions in decoding tasks. These findings underscore EEG2GAIT's potential for advancing brain-computer interface applications, particularly in lower-limb rehabilitation and assistive technologies.

Xi Fu, Weibang Jiang, Rui Liu et al.

Accurate decoding of lower-limb motion from EEG signals is essential for advancing brain-computer interface (BCI) applications in movement intent recognition and control. However, challenges persist in achieving causal, phase-consistent predictions and in modeling both inter- and intra-subject variability. To address these issues, we propose NeuroDyGait, a domain-generalizable EEG-to-motion decoding framework that leverages structured contrastive representation learning and relational domain modeling. The proposed method employs relative contrastive learning to achieve semantic alignment between EEG and motion embeddings. Furthermore, a multi-cycle gait reconstruction objective is introduced to enforce temporal coherence and maintain biomechanical consistency. To promote inter-session generalization, during fine-tuning, a domain dynamic decoding mechanism adaptively assigns session-specific prediction heads and learns to mix their outputs based on inter-session relationships. NeuroDyGait enables zero-shot motion prediction for unseen individuals without requiring adaptation and achieves superior performance in cross-subject gait decoding on benchmark datasets. Additionally, it demonstrates strong phase-detection capabilities even without explicit phase supervision during training. These findings highlight the potential of relational domain learning in enabling scalable, target-free deployment of BCIs.