The trajectory planning problem is a key aspect of underwater robot operations. Most existing studies focus on a single underwater robot and environmental obstacles, with limited research on collision avoidance planning among multiple underwater robots. This paper addresses these issues for multiple underwater robots by performing the simultaneous target tracking (STT) from an optimization perspective. On the basis of the neurodynamic design method, the path tracking, collision avoidance, and communication maintenance are formulated as a series of inequality and equality constraints. By incorporating these constraints, the STT scheme, which is depicted as a unified quadratic program (QP) framework, is thus established for multiple underwater robots with certified safety. Such a QP-type STT scheme is then solved by using the Lagrangian-based recurrent neural network. Computer simulations are performed to further verify the effectiveness and feasibility of the proposed QP-type STT scheme. These results indicate that the multiple underwater robots via the proposed STT scheme can accomplish the assigned target-tracking tasks while maintaining the safe distance to avoid collision and staying within communication range throughout the tracking mission.

With the rise of virtual communication and smart devices, speech has become the most natural medium of interaction. Yet it remains intrinsically difficult: speech is fleeting, unstructured, and disfluent, making key information prone to loss. Conventional Speech-to-Text (STT) systems attempt to acoustically reconstruct what was said. However, their frame-level alignment and rigid token-by-token decoding break down under noise, interruptions, or fragmentation. Humans, in contrast, readily grasp what was meant by exploiting syntax, discourse, pragmatics, and prosody. We argue for a paradigm shift from acoustic reconstruction to semantic transduction: inferring meaning directly from speech, abstracted from surface distortions. This shift raises two challenges: (C1) the lack of anchors between audio and meaning, and (C2) the need to maintain compositional semantics. To address these, we introduce CogTrans, a cognitively inspired speech-to-meaning framework. CogTrans tackles C1 through a Semantic Anchor Explorer, built on I-JEPA to capture higher-order regularities, prosodic rhythms, cross-frequency coarticulation, discourse continuity-providing resilient semantic scaffolds under noise and fragmentation. For C2, it designs a Lexical-Semantic Harmonizer that dynamically integrates these anchors with lexical embeddings; thereby preserving fine-grained compositional fidelity in roles, order, and entities. Extensive experiments show that CogTrans delivers consistent and substantial gains under challenging conditions. On GigaSpeech, it achieves a 6.58% relative Word Error Rate (WER) reduction, and on the multilingual VoxPopuli benchmark, the gain climbs to 12.97% at 10 dB noise-a regime where conventional models typically collapse. Beyond literal accuracy, CogTrans also boosts semantic fidelity, with a 3.40% increase in ROUGE-L and 3.45% in USE-Sim, ensuring transcripts remain faithful not only in words but also in meaning. Together, these results underscore that CogTrans is robust in noisy, unconstrained environments-precisely the conditions where reliability matters most.

We introduce an interactive music system powered by AI-generated content (AIGC) that enables users to engage with music through multimodal interactions involving touch, sound, and spatial immersion. Motivated by the desire to enhance engagement and emotional connection with music, our system enables users to co-create and interact with musical content. Users upload a song and a descriptive text prompt, from which the system generates 3D visuals. During playback, users can embed their own audio inputs and trigger responsive visual effects such as color-driven point clouds using tangible controls. To explore how spatial scale and embodiment shape user experience, we implement the system across three increasing spatial scales and embodiments: (1) a handheld AR music box, (2) a table-sized stage box, and (3) a fully immersive VR environment. Through a user study, we investigate how different levels of immersion and interaction influence user engagement, emotional response, and sense of presence. Our findings demonstrate the potential of combining AIGC with embodied interaction to enrich creative expression and enhance immersive musical experiences.

This study examines the effects of two visual guidance techniques, Visual Enhancement and Visual Suppression, on user perception of contextual information in video content. Visual Enhancement introduces explicit visual cues to highlight target content, whereas Visual Suppression attenuates non-target elements, for example, by reducing their brightness. Both approaches aim to isolate specific objects from the background, directing attention to critical information within complex, dynamic scenes. Despite their growing usage, the relative effectiveness of these approaches in guiding attention and their impact on peripheral context awareness remain underexplored. To address this gap, we conducted a controlled user study with 27 participants. The results indicate that Visual Enhancement, through the addition of salient cues, more effectively directs user attention to target information than Visual Suppression. Our findings advance understanding of visual attention in dynamic environments and offer implications for designing visual guidance strategies.

In the architecture of Digital Twin(DT) optical networks, model management is a key component to ensure efficient network operations. It involves not only precise digital mapping of physical network entities but also includes model integration, dynamic updates, and real-time inference. Considering the diversity and complexity of model types when building an optical network DT, this paper proposes a novel DT model management framework for Optical Transport Network (OTN). This framework enables uploading, training, evaluating, and deploying models to perform inference tasks. It supports both online and offline training to meet diverse business demands. Furthermore, to adapt to continuously changing data and environments, the system integrates an incremental learning technique, which enables continuous updates and optimization of models. Additionally, the framework supports the deployment and inference of mathematical models. Additionally, the components deployed by this framework are entirely managed through containerization technology. This approach ensures portability and enhances the system's fault tolerance and high availability. This framework provides an efficient, user-friendly, and scalable solution for comprehensive model management.

The existing masked face recognition algorithms almost tend to adopt synthetic masked face datasets for training. However, these models are limited as they rely on existing mask augmentation methods, which contain few mask patterns and cannot simulate shadows and textures in realistic scenes. To overcome this limitation, we propose a semi-supervised face recognition framework to fully exploit unlabeled real masked face samples, improving the mask robustness of the recognition model. More specifically, unlike the original face embedding network, we design a part-aware network to explore multi-region face representation based on the face structure. In this way, we obtain multiple face sub-embeddings, which correspond to different regions of the face, including the upper half, the lower half and the whole. Crucially, we use the norm of the sub-embedding to represent the activation state of the facial region features. For the input unlabeled masked face image, we restrict the sub-embedding norm of its lower half to weaken the face feature representation of the occluded area. For normal face samples, their partial features are kept activated by maintaining the sub-embedding norm, which guides the deep network does not ignore the available information. Moreover, we employ the margin-based recognition loss for normal samples to ensure that the model is sufficiently discriminative for normal facial features. Extensive experimental results on both normal and real masked face datasets show that our approach significantly outperforms the state-of-the-arts. Code is available at https://github.com/Baojin-Huang/UFace.

No-reference image quality assessment (NR-IQA) methods, inspired by the free energy principle, improve the accuracy of image quality prediction by simulating the human brain’s repair process for distorted images. However, existing methods use separate optimization schemes for distortion restoration and quality prediction, which undermines the accurate mapping of feature representations to quality scores. To address this issue, we propose a joint restoration and quality feature learning NR-IQA (RQFL-IQA) method to jointly tackle distortion image restoration and quality prediction within a unified framework. To accurately establish the quality reconstruction relationship between distorted and restored images, a hybrid loss function based on pixel-wise and structure-wise representations is used to improve the restoration capability of the image restoration network. The proposed RQFL-IQA exploits rich labels, including restored images and quality scores, to enable the model to learn more discriminative features and establish a more accurate mapping from feature representation to quality scores. In addition, to avoid the impact of poor restoration on quality prediction, we propose a module with a cleaning function to reweight the fusion of restored and primitive features to achieve more perceptual consistency in feature fusion. Experimental results on public IQA datasets show that the proposed RQFL-IQA is superior over existing methods.

Image quality assessment (IQA) is an important problem in computer vision with many applications. We propose a transformer-based multi-task learning framework for the IQA task. Two subtasks: constructing an auxiliary information error map and completing image quality prediction, are jointly optimized using a shared feature extractor. We use visual transformers (ViT) as a feature extractor for feature extraction and guide ViT to focus on image quality-related features by building auxiliary information error map subtask. In particular, we propose a fusion network that includes a channel focus module. Unlike the fusion methods commonly used in previous IQA methods, we use the fusion network, including the channel attention module, to fuse the auxiliary information error map features with the image features, which facilitates the model to mine the image quality features for more accurate image quality assessment. And by jointly optimizing the two subtasks, ViT focuses more on extracting image quality features and building a more precise mapping from feature representation to quality score. With slight adjustments to the model, our approach can be used in both no-reference (NR) and full-reference (FR) IQA environments. We evaluate the proposed method in multiple IQA databases, showing better performance than state-of-the-art FR and NR IQA methods.

In this paper, we consider the face swapping detection from the perspective of face identity. Face swapping aims to replace the target face with the source face and generate the fake face that the human cannot distinguish between real and fake. We argue that the fake face contains the explicit identity and implicit identity, which respectively corresponds to the identity of the source face and target face during face swapping. Note that the explicit identities of faces can be extracted by regular face recognizers. Particularly, the implicit identity of real face is consistent with the its explicit identity. Thus the difference between explicit and implicit identity of face facilitates face swapping detection. Following this idea, we propose a novel implicit identity driven framework for face swapping detection. Specifically, we design an explicit identity contrast (EIC) loss and an implicit identity exploration (IIE) loss, which supervises a CNN backbone to embed face images into the implicit identity space. Under the guidance of EIC, real samples are pulled closer to their explicit identities, while fake samples are pushed away from their explicit identities. More-over, IIE is derived from the margin-based classification loss function, which encourages the fake faces with known target identities to enjoy intra-class compactness and inter-class diversity. Extensive experiments and visualizations on several datasets demonstrate the generalization of our method against the state-of-the-art counterparts.

Most existing blind image quality assessment (BIQA) models focus on improving performance on existing datasets and are weak in adapting to unknown distortion or degradation types. In this paper, we propose a Transformer-based BIQA contrastive continual learning approach to improve model transfer performance. The basic idea is that the model continuously learns from the IQA data stream, integrating new knowledge from the current dataset. At the same time, limited access to previous data using a limited memory budget prevents forgetting the knowledge gained from the dataset of old tasks. We design an attentional contrastive learning strategy based on the Transformer architecture with a designed attentional focus contrastive loss to rebalance the contrastive learning between the new and old tasks, which can consolidate the previously learned representations. In addition, we used a structure similar to the cumulative classifier, balancing the learning of the current task quality score with the quality scores of all observed tasks. Extensive experiments demonstrate the feasibility of the proposed continuous learning approach compared to the standard training techniques of BIQA.

Mutation profiles together with prior knowledge such as interactions between genes/proteins provide abundant critical information for the identification of driver modules, which is very important for analyzing mutational heterogeneity in human cancers. Due to the negative effects of inevitable false positive interactions in the PPI network, subcellular localization data are exerted to filter out them firstly, and somatic mutation profiles are used to weight the retained interactions. Five novel recombination operators are introduced basing on the vertex degrees and the edge weights in the PPI network, and a parthenogenetic algorithm is devised for solving the presented identification model which takes into account network connectivity, mutual exclusivity, coverage, and hops between genes within a module. Extensive experimental results indicate that compared with two state-of-the-art computational methods Hotnet2 and MEXCOwalk, the proposed method exhibits competitive performance in most cases in terms of recovering known cancer genes, providing modules that have satisfied coverage and mutual exclusivity, and are enriched for mutations in specific cancer types. Many identified gene sets are involved in known signaling pathways, most of the implicated genes are oncogenes or tumor suppressors previously reported in the literature. In addition, the proposed method does identify many cancer related genes missed by methods Hotnet2 and MEXCOwalk, including some recognized genes covering many types of cancers but having low mutation frequency.