Qingdao Academy of Intelligent Industries

RGB-D salient object detection (RGB-D SOD) aims to accurately localize and segment visually salient objects by jointly leveraging RGB images and depth maps. Some existing methods rely on static fusion strategies with fixed paths and weights, which treat all regions equally and fail to capture the varying importance of different regions and modalities. Although some attention-based methods alleviate the limitations of static fusion by assigning adaptive weights to different regions and modalities, the quality of RGB and depth data may degrade in real-world scenarios due to sensor noise, illumination changes, or environmental interference. These attention-based methods often overlook inter-modality quality differences and complementarity, making them prone to over-relying on a certain modality, which can lead to noise introduction, feature conflicts, and performance degradation. To address these limitations, this paper proposes a novel dynamic feature routing and fusion framework for RGB-D SOD, which adaptively adjusts the fusion strategy according to the quality of input modalities. To enable modality quality awareness, the proposed method characterizes the modality complementarity between RGB and depth features in a task-driven manner inspired by information-theoretic principles. We introduce a task-relevance scoring function which is integrated with a mutual information estimator to quantify such complementarity, and emphasizes task-relevant features while suppressing redundancy. A dynamic routing module is then designed to perform feature selection guided by the captured complementarity. In addition, we propose a novel cross-modal fusion module to adaptively fuse the features selected by the dynamic routing module, which effectively enhances complementary representations while suppressing redundant features and noise interference. Extensive experiments conducted on seven public RGB-D SOD benchmark datasets demonstrate that the proposed method consistently achieves competitive performance, outperforming existing methods by an average of approximately 1% across multiple evaluation metrics. Notably, in challenging scenarios with severe modality quality degradation, the proposed method outperforms existing best-performing methods by up to 1.8%, demonstrating strong robustness against cluttered backgrounds, complex object structures, and diverse object scales. Overall, the proposed dynamic fusion framework provides a novel solution to modality quality imbalance in RGB-D salient object detection.

Salient object ranking aims to assign a relative importance order to multiple objects in an image, aligning with human visual attention. However, existing methods struggle with ranking ambiguity in complex scenes, particularly when objects are numerous, occluded, or semantically similar, leading to decreased accuracy for low-saliency objects. To address this, we propose PairwiseSOR-MLMs, a novel framework leveraging multimodal large models and pairwise comparison to achieve salient object ranking. The approach decomposes global ranking into a series of pairwise comparison tasks. It first employs object detection and instance segmentation to identify objects, uses image inpainting to reconstruct scenes by removing occlusions, and then prompts MLMs to perform pairwise comparisons based on visual saliency cues. Finally, another MLM inference aggregates these comparisons into a consistent global ranking. Experiments on ASSR and IRSR benchmarks show our method achieves state-of-the-art or competitive performance across metrics, demonstrating robustness in handling occlusion and semantic similarity. Its pairwise comparison paradigm can extend to other relative assessment tasks.

To achieve high-quality microhole processing at elevated laser powers, this work proposes a synergistic strategy—backside flowing chemical-assisted laser pulse-delayed scanning drilling [time-delayed (TD)-backside chemical-assisted laser drilling (BCALD)]. Unlike conventional immersion-based liquid-assisted laser drilling, which often suffers from significant plasma shielding and uncontrollable chemical etching, TD-BCALD significantly enhances energy utilization efficiency by precisely modulating the spatiotemporal coupling between laser pulses and chemical reactions. By coordinating a continuous flow of environmentally benign chemical fluid with an intermittently activated high-power laser in a pulse-delay regime, this approach enables highly controllable synergistic interactions within the laser-induced high-temperature zone. We first outline the individual mechanisms and key parameters of BCALD and laser pulse-delayed scanning drilling and then highlight how pulse-delay timing critically modulates their synergy. Detailed scanning electron microscopy and x-ray energy spectrometry analyses further elucidate the underlying mechanism of heat accumulation decoupling. Under an optimal 100 ms pulse delay, TD-BCALD achieves a taper angle of 0.987°, roundness of 98.9%, and surface roughness of 1.502 μm. These results, combined with the reduced chemical consumption and improved process stability, demonstrate the superior engineering potential and environmental benefits of TD-BCALD for advanced high-power laser microfabrication.

Electrostatic field-driven minimum quantity lubrication (EMQL) demonstrates unique advantages in enhancing lubrication and cooling performance of droplet. However, while considerable research has focused on evaluating the machining performance of EMQL, the underlying charging mechanisms and atomization dynamics of bio-based lubricants remain unclear. This study introduces a novel charging nozzle that employs both contact and corona charging mechanisms. It then investigates the charging mechanisms of biolubricants based on electric field distribution characteristics. Furthermore, the charging, atomization, and spreading performance of various biolubricants are experimentally evaluated. The results demonstrate that the combined contact-corona charging nozzle significantly enhances surface charge density in both droplets and the cutting zone, compared to conventional contact charging nozzles. Biolubricants with high electric conductivity additives show superior charging properties over pure vegetable oil. Increasing the applied voltage from 0 to 35 kV leads to substantial reductions in mean droplet diameter: 28.83% for rapeseed oil, 38.94% for lecithin-oil mixture, 35.76% for WS 2 nanofluids, and 38.28% for CNT nanofluids. Atomization dynamics analysis reveals that the presence of charges on droplet surfaces and the electric field at the nozzle both contribute to enhancing the atomization performance of biolubricants. However, higher conductivity accelerates interfacial charge release, which inhibits electrically driven droplet spreading. This study provides valuable insights into the EMQL mechanism through experimental and theoretical analysis, expanding the potential applications of charged biolubricants in clean manufacturing and tribology.

ABSTRACT As a highly promising tool, femtosecond lasers are anticipated to become a powerful engine and support for the advancement of frontier science and advanced intelligent manufacturing technologies. This paper explores the fabrication techniques of femtosecond laser micro‐nano structures and their functional applications, unveiling the theoretical models and processing mechanisms of ultrafast laser machining, and detailing the extensive applications in surface modification, laser patterning, and two‐photon polymerization. In addressing the technical challenges of ultrafast lasers related to processing efficiency and accuracy, the footprints of femtosecond laser high‐efficiency precision manufacturing technologies were traced and investigated. Subsequently, the technological characteristics of holographic laser processing are elucidated, creating favorable conditions for the digitization and industrialization of femtosecond laser precision manufacturing with high flexibility, high quality, and high efficiency. The development trajectory and future prospects of holographic femtosecond laser beam shaping in two‐dimensional (2D) and three‐dimensional (3D) processing technologies are also summarized. The findings of this research will provide valuable insights and serve as a reference for forward‐looking and strategic innovation in advanced manufacturing fields such as aerospace, integrated circuits, information processing, and biomedicine.

Graph Neural Networks (GNNs) have become foundational models in recommender systems due to their ability to propagate information over user–item bipartite graphs via neighborhood aggregation. Despite their empirical success, GNNs are inherently constrained by their reliance on local connectivity, which limits their ability to capture global interaction patterns, particularly in large-scale recommendation scenarios characterized by severe data sparsity. To address these challenges, we propose the Taylor Linear attention in Transformer (TLFormer), which enhances recommendation performance by enabling global attention across all user–item pairs while preserving graph structural information. Unlike existing Transformer-based recommendation approaches that focus on local attention patterns, TLFormer introduces a novel linear attention mechanism derived from the first-order Taylor approximation, allowing efficient computation of all-pair interactions. TLFormer integrates spatial topology as positional encoding while maintaining linear complexity, effectively balancing computational efficiency with model expressiveness for large-scale recommendation scenarios. Extensive experiments across multiple datasets demonstrate that TLFormer significantly outperforms state-of-the-art methods, particularly in scenarios with sparse interactions and long-tail distributions.

Abstract. As the largest oligotrophic ocean globally, the North Pacific oligotrophic ocean gyre (NPOG) exhibits pronounced variability on seasonal, decadal, and centennial time scales. Notably, changes in the seasonality of the NPOG are thought to have larger effects on marine ecosystems than changes in its annual mean state. However, the interannual variability of NPOG seasonality and its response to climate processes remain unclear. Here, we investigate the amplitude of the seasonal cycle in NPOG area and its linkage with climate variability and change. Our results show that the El Niño–Southern Oscillation (ENSO) modulated the seasonal maximum of NPOG area in boreal summer, and thus the amplitude of the seasonal cycle during 1998–2021. This is primarily due to ENSO-induced changes in nutrient transport via equatorial upwelling and thermal stratification, as well as changes in the chlorophyll-to-carbon ratio in phytoplankton cells (photoacclimation). Future projections based on Coupled Model Intercomparison Project Phase 5 (CMIP5) modelling results and an Elman neural network indicate a significant decrease in the seasonal amplitude of NPOG area by 2100, attributed to the growing seasonal minimum of NPOG area in winter along the anthropogenic increase in atmospheric CO2. The findings highlight the importance of considering seasonal differences in future research on the interannual variability of oligotrophic gyres and underscore the need for models to distinguish between the effects of climate variability and change.