Zhijing Xu

Galaxy morphology detection is a pivotal task for unraveling cosmic evolutionary mechanisms, yet existing models exhibit insufficient detection accuracy for irregular and small-target galaxies. To address this, this paper proposes the STAR-YOLO galaxy morphology detection model. The backbone network incorporates the novel Multi-scale Attentive Context Aggregation module, which deeply integrates multi-scale dilated convolution with a progressive spatial-channel attention mechanism to enhance feature extraction for irregular and small galaxies. Meanwhile, we design the lightweight Lightweight Efficient Attention Network module that reduces parameters through channel compression. The proposed Adaptive Focal Spatial-IoU loss function further improves detection performance for small galaxies through dynamic focal mechanisms and scale-invariant optimization. Evaluated on Galaxy Zoo 2 data set, our STAR-YOLO achieves 96.3% mean average precision—a 2.5% improvement over baseline models, with irregular galaxy recognition accuracy notably increasing by 9.3%. Comparative experiments demonstrate superior detection capabilities for multi-target irregular galaxies compared to state-of-the-art models, providing an innovative solution for astronomical image analysis.

This study introduces Ship-Yolo, a novel algorithm designed to tackle the challenges of detecting small targets against complex backgrounds in remote sensing imagery. Firstly, the proposed method integrates an efficient local attention mechanism into the C3 module of the neck network, forming the EDC module. This enhancement significantly improves the model’s capability to capture critical features, enabling robust performance in scenarios involving intricate backgrounds and multi-scale targets. Secondly, a Lightweight Asymmetric Decoupled Head (LADH-Head) is proposed to separate classification and regression tasks, reducing task conflicts, improving detection performance, and maintaining the model’s lightweight characteristics. Additionally, the LiteConv module is designed to replace the C3 module in the backbone network, leveraging partial convolution to ignore invalid information in occluded regions and avoid misjudgments. Finally, the Content-Aware Reassembly Upsampling Module (CARAFE) is employed to replace the original upsampling module, expanding the receptive field to better capture global information while preserving the lightweight nature of the model. Experiments on the ShipRSImageNet and DOTA datasets demonstrate that Ship-Yolo outperforms other YOLO variants and existing methods in terms of precision, recall, and average precision, exhibiting strong generalization capabilities. Ablation studies further validate the stable performance improvements contributed by the EDC, LADH-Head, LiteConv, and CARAFE modules.

The prediction of gas pipeline leakage is not only of great significance to public safety and environmental protection but also plays a crucial role in economic benefits, management efficiency, and social responsibility. This paper introduces an improved Transformer-based method for gas pipeline leakage prediction. This method leverages the unique advantages of the Transformer model's self-attention mechanism, enabling efficient handling of dependencies in long time series and capturing complex relationships between multidimensional data. The proposed improvements enhance the Transformer model's generalization ability, robustness, and accuracy when dealing with multivariate time series data. During training, adversarial samples were generated for adversarial training to increase the model's robustness against minor perturbations; structured anomalies were added to the data, and regularization loss was calculated to prevent overfitting. Experiments based on real multidimensional sensor data demonstrate that the proposed model significantly improves the accuracy of gas pipeline leakage prediction.

The purpose of this research is to tackle one of the most difficult issues in the realm of self-driving cars, which is the testing of advanced self-driving application scenarios. Thus, this study proposes a simulation testing system based on hardware-in-the-loop simulation technology. This system can enable data exchange between hardware systems as well as replicate and evaluate the algorithmic operations of the equipment under laboratory conditions. The system can integrate scenario simulation software with MATLAB to evaluate algorithm performance. The vehicle formation control system is tailored for collaborative vehicle formation management scenarios and tested in the simulation test system. The findings display the functional integrity of the vehicle formation control system, the reliability of lane changing and the stability and safety of cruising. It additionally demonstrates that the simulation testing system has the ability to recreate cooperative vehicle arrangement management situations and assess their functionality and performance. In forthcoming research, comprehensive functional and performance assessments will be executed on various typical scenarios for advanced autonomous driving applications in order to authenticate the simulation test system’s applicability.

Humans express their emotions in a variety of ways, which inspires research on multimodal fusion‐based emotion recognition that utilizes different modalities to achieve information complementation. However, extracting deep emotional features from different modalities and fusing them remain a challenging task. It is essential to exploit the advantages of different extraction and fusion approaches to capture the emotional information contained within and across modalities. In this paper, we present a novel multimodal emotion recognition framework called multimodal emotion recognition based on cascaded multichannel and hierarchical fusion (CMC‐HF), where visual, speech, and text signals are simultaneously utilized as multimodal inputs. First, three cascaded channels based on deep learning technology perform feature extraction for the three modalities separately to enhance deeper information extraction ability within each modality and improve recognition performance. Second, an improved hierarchical fusion module is introduced to promote intermodality interactions of three modalities and further improve recognition and classification accuracy. Finally, to validate the effectiveness of the designed CMC‐HF model, some experiments are conducted to evaluate two benchmark datasets, IEMOCAP and CMU‐MOSI. The results show that we achieved an almost 2%∼3.2% increase in accuracy of the four classes for the IEMOCAP dataset as well as an improvement of 0.9%∼2.5% in the average class accuracy for the CMU‐MOSI dataset when compared to the existing state‐of‐the‐art methods. The ablation experimental results indicate that the cascaded feature extraction method and the hierarchical fusion method make a significant contribution to multimodal emotion recognition, suggesting that the three modalities contain deeper information interactions of both intermodality and intramodality. Hence, the proposed model has better overall performance and achieves higher recognition efficiency and better robustness.