Tongue diagnosis is integral to Traditional Chinese Medicine (TCM) for evaluating a patient’s body constitution. Yet, this field faces challenges such as indirect constitution diagnosis, a dearth of labeled datasets, and the complexities of few-shot learning. Existing studies focus mainly on analyzing tongue color and coating, rather than directly analyzing the body constitution from the patient’s tongue. Moreover, the lack of publicly available datasets with constitution labels impedes model development for constitution diagnosis. The resource-intensive process of creating large-scale labeled datasets calls for efficient training methods on small datasets. Addressing these issues, this paper presents an In-Context Multitask Learning approach to improve few-shot fine-tuning accuracy for Large Language Models (LLMs) in constitution diagnosis. We created a dataset for color-coating analysis and a few-shot dataset with constitution labels and a structured prompt-label framework, allowing LLMs to learn from varied datasets. Our method shows improved performance over traditional methods in constitution diagnosis and significantly boosts LLMs’ generalization and robustness in TCM’s complex diagnostic tasks, providing a viable path for automating tongue diagnosis.

The prevalence of depression is escalating, especially among youth, which has become a critical mental health concern. Current assessment methods, relying heavily on questionnaires, clinical observations, and AI-driven analyses, are limited by their focus on single-event data, failing to encapsulate the nuanced expressions of depressive symptoms. Moreover, a significant oversight in existing research is the underutilization of electromyogram (EMG) alongside electroencephalogram (EEG) data, which could provide a more holistic view of unconscious body behaviors. Additionally, given the high variability of depression among individuals, traditional analysis models are in urgent need of refinement to accommodate the personality of different individuals. To address these limitations, we propose M3ADD, a novel benchmark for Automatic Depression Detection that employs a Multimodal, Multitask, and Multievent framework. We collected EEG and EMG data from 97 participants across varied events (interview, reading tasks, walking), coupled with standardized questionnaires assessing depression, wellbeing, and personality, enriching our multitask learning approach. Our benchmark recognition algorithm leverages multitask learning, channel and interactive attention mechanisms to synthesize event-specific and modal-specific features, enhancing adaptability to individual differences and improving data utilization efficiency. M3ADD surpasses existing models by achieving 87% accuracy in detecting depression and 95% accuracy in assessing wellbeing, providing a promising avenue for early identification.

The conversion of monocular RGB images to depth maps is crucial in robotic applications. Current supervised learning approaches, dependent on high-quality RGB-Depth pairs, struggle with indoor environments characterized by multiple objects and fluctuating lighting, leading to inaccurate and unstable depth estimations. Additionally, commercial depth cameras on robots produce incomplete and light-sensitive depth maps, further degrading estimation quality. To surmount these issues, We created a comprehensive dataset of 9,600 RGB-Depth image pairs, capturing a range of indoor scenes under various lighting (dim, normal, and strong lighting), and interference conditions (local strong light interference, specular reflection interference, background similarity interference, and combinations of these factors). This dataset serves as a foundation for our proposed monocular RGB-to-depth mapping framework, which employs a hierarchical similarity loss to enhance the model’s structural feature learning from reference depth maps, improving the fidelity of estimated depth maps. We also integrated adversarial training and attention mechanisms to refine the depth consistency between the estimated and original depth maps. Experiments show our method surpasses current benchmarks in metrics like Threshold Accuracy, SILog, and MSE, demonstrating its robustness and potential for real-world applications, even with poor-quality inputs.

Depression is a widespread mental health issue affecting diverse age groups, with notable prevalence among college students and the elderly. However, existing datasets and detection methods primarily focus on young adults, neglecting the broader age spectrum and individual differences that influence depression manifestation. Current approaches often establish a direct mapping between multimodal data and depression indicators, failing to capture the complexity and diversity of depression across individuals. This challenge includes two tracks based on age-specific subsets: Track 1 uses the MPDD-Elderly dataset for detecting depression in older adults, and Track 2 uses the MPDD-Young dataset for detecting depression in younger participants. The Multimodal Personality-aware Depression Detection (MPDD) Challenge aims to address this gap by incorporating multimodal data alongside individual difference factors. We provide a baseline model that fuses audio and video modalities with individual difference information to detect depression manifestations in diverse populations. This challenge aims to promote the development of more personalized and accurate de pression detection methods, advancing mental health research and fostering inclusive detection systems. More details are available on the official challenge website: https://hacilab.github.io/MPDDChallenge.github.io.

Intelligent sensing technologies are transforming sports training by enabling precise motion analysis, critical for skill development and performance optimization. This study introduces a badminton racket handle embedded with a lightweight, multi-node MEMS-based sensing system designed for real-time motion recognition. To capture distributed grip forces, swing trajectories, and impact mechanics at the player-equipment interface, the system employs an ergonomic design ensuring natural gameplay. A hybrid feature extraction approach, integrating time-and frequency-domain features with a 1D-CNN, achieves a classification accuracy of 97.89% across ten badminton actions. To enhance interpretability and provide actionable insights, explainable AI using SMDL-attribution identifies key motion features, revealing biomechanical inefficiencies in grip strength, swing consistency, and wrist motion. Seamlessly integrated with Virtual Reality (VR) platforms, the system delivers immersive, real-time feedback, transforming training into an interactive and data-driven experience. By combining advanced sensing, machine learning, and explainable AI, this system establishes a new benchmark for intelligent sports monitoring, with broad applications in sports training, rehabilitation, and human-computer interaction.

Motion recognition in IoT-based sensor systems is crucial for applications such as healthcare and human-computer interaction. However, one key challenge—data structure inconsistencies—complicates the performance of existing systems, particularly in dynamic real-world environments. Traditional fusion approaches lack the adaptability required to address sensor inconsistencies and fail to fully leverage the potential of multisensor data. To overcome this challenge, we propose a dynamic multisensor fusion framework (DMSFF) with adaptive spatiotemporal optimization. This framework introduces a dynamic sensor weighting mechanism that prioritizes reliable data while suppressing noise, ensuring robustnesss. A transformer-based fusion architecture captures spatiotemporal features, modeling complex intersensor relationships and long-term dependencies. Additionally, a motion kernel matching module aligns the data with canonical motion patterns, improving feature extraction and enhancing the recognition of subtle activities. The framework is validated on benchmark datasets, including those with real-world noise and structural inconsistencies, achieving an accuracy of 99.48%. This work establishes a new benchmark for multisensor motion recognition, providing scalable and robust solutions for smart healthcare and human-computer interaction.

Recent advances in personality recognition have improved trait inference from multimodal data, yet many existing methods rely on short-term video segments or static images, limiting the modeling of temporal dynamics due to short video durations, sparse frame-level annotations, and inconsistent modality coverage across audio, text, and visual channels. These limitations make it difficult to model how personality traits manifest over time and across modalities in naturalistic settings. To address these challenges, we introduce the Northeast University Personality Recognition (NEUPR) dataset, comprising 654 self-reported and discussion-based videos collected through MBTI assessments. NEUPR offers naturally expressed multimodal dataincluding audio, facial expressions, eye movements, and speech transcripts-captured across diverse participants and real-world settings. Building on this dataset, we propose MPRNet, a unified framework for dynamic personality recognition featuring two core innovations: (1) a multimodal encoder that leverages LSTM to capture temporal dependencies across longer sequences and integrates latent personality embeddings extracted from BERT representations of text to enrich semantic context, fused through adaptive weighting and enhanced by Gram encoding to preserve local feature patterns; and (2) a feature enhancement module that incorporates learnable positional encoding and channel attention to address modality imbalance and improve sensitivity to spatially salient features across modalities. Experimental results demonstrate that MPRNet outperforms state-of-the-art methods across multiple datasets, while ablation studies confirm the effectiveness of its components. By explicitly modeling temporal variation and enhancing cross-modal fusion, MPRNet enables more robust personality inference. This work establishes both a benchmark dataset and an adaptive modeling framework for multimodal personality analysis, advancing dynamic trait recognition.

Depression is prevalent in patients with Parkinson's disease (PD), due to the dramatic negative impact that behavioral disorders have on daily life. Regrettably, most researchers in the past ignored the study of depression in PD patients, especially when depressive symptoms and PD symptoms are coupled together, it is difficult for researchers to recognize depression from the macro physiological signs of PD patients. Researchers are increasingly turning their attention to the subtle phenomena of emotional expression in conversation, using the textual and spectral features extracted from the audio of interviews as the primary support for understanding emotional states. However, there is still a lack of effective technical means to fuse these two features to recognize depression in PD patients. In this study, we proposed an innovative image fusion approach, fusing a sparse emotion dictionary with textual features and a Continuous Wavelet Transform (CWT) spectrogram with spectral features for the precise recognition of depression in PD patients. The fusion process integrates low-dimensional emotion-related textual cues, contributing to a more comprehensive extraction of emotionally relevant information. Subsequently, we introduce a High and Low Frequency Feature Fusion Multi-headed Self-Attention (HL-MSA) mechanism within a high and low frequency feature fusion network to amalgamate information across different frequency features within the images. The results underscore the efficacy of this novel fusion approach in effectively extracting depressive features in PD patients, attaining advanced recognition performance. Notably, this endeavor represents a pioneering stride in seamlessly fusing a sparse emotion dictionary and CWT spectrogram, exemplifying a promising and effective initiative for recognizing depression in PD patients.

The diagnosis of depression in individuals with Parkinson's Disease (PD) through the utilization of multimodal fusion techniques represents a significant domain. The primary challenge involves the creation of a robust fusion framework to address the heterogeneity among different modalities effectively. However, previous studies primarily focused on interactions between heterogeneous data, neglecting the structural similarities among isomorphic data, resulting in a substantial loss of feature information when merging heterogeneous data. In this study, we introduced a multi-modal data image encoding and fusion approach for diagnosing depression in PD patients. Additionally, we proposed a multi-modal dataset encompassing motion, facial expression, and audio data. First, we designed an RGB and sparse coding method to encode the multi-modal data, achieving the isomorphic transformation of multi-modal information and extracting feature information from lower-dimensional spaces. Furthermore, we introduced a Spatial-Temporal Network (STN) to fuse the three types of encoded images. We incorporated the Relation Global Attention (RGA) to enhance feature extraction and leverage all encoded image location feature nodes for balanced decision attention. Finally, recognizing the limitations of traditional machine learning algorithms in handling multi-tasks in medical diagnosis, we established a multi-task weighted loss function to achieve depression identification and severity prediction through Multi-Task learning (MTL).

Elderly individuals often suffer from underlying medical conditions, resulting in a significant decline in quality of life and a heightened susceptibility to depression. Presently, AI screening tools based on behavioral indicators offer an objective and effective approach to diagnosing depression. However, current AI depression screening tools are primarily tailored to adolescents and adults, exhibiting shortcomings in their applicability and accuracy for elderly individuals with underlying medical conditions. To address the above issues, first, this paper constructs a depression dataset for elderly people with underlying diseases by using semi-structured interviews. Second, based on cognitive science insights, it is recognized that personality factors significantly influence behavioral expressions and also determine the attitudes of elderly individuals toward current life circumstances/health issues. Therefore, besides annotating depression severity, the Big Five-10 personality scale was utilized to annotate participant personalities. Finally, a late fusion-based multi-task learning framework was proposed, and the effects of introducing gait information and personality annotation on the performance of depression assessment were investigated. The experimental findings affirm the importance of integrating gait information and personality assessment in improving depression detection effectiveness. This study provides valuable foundational resources, as well as beneficial references and insights, for the research on depression in the elderly.

The detection and statistics of damaged buildings are crucial for rescue work, especially when detecting object-level and small buildings. Existing detection technologies mainly focus on pixel-level analysis, often ignoring the significance of object-level analysis and the identification of small buildings. To address these challenges, a Cross-Temporal Progressive Enhancement Model (CTPEM) is proposed. In CTPEM, the Visual Center Guided Enhancement (VCGE) module is proposed, which uses multiple visual centers to learn the binary relationship between pre- and post-temporal features to capture global and local information. Meanwhile, the Progressive Small Object Enhancement (PSOE) module is proposed, which is used to capture tiny and small-sized features through multiple feature maps at varying depths, aiming to reduce the influence of gradually vanishing features. Compared with state-of-the-art (SOTA) methods, CTPEM achieves superior performance in detecting damaged buildings, particularly small and tiny targets. Extensive experimental results demonstrate the effectiveness and practical value of our approach for accurate casualty assessment, and further facilitate more efficient disaster response and optimal resource allocation. The CTPEM code is available at https://github.com/GZPLHJ181107/CTPEM.

The outbreak of COVID-19 has greatly changed everyone's lifestyle all over the world. One of the best ways to prevent the spread of infections is by washing hands properly. Although a number of hand hygiene monitoring systems have been proposed, they either cannot achieve high accuracy in practice or work only in limited environments such as hospitals. Therefore, a ubiquitous, energy-efficient and highly accurate hand hygiene monitoring system is still lacking. In this paper, we presentWashRing—the first smart ring-based handwashing monitoring system. In WashRing, we design a Partially Observable Markov Decision Process (POMDP) based adaptive sampling approach to achieve high energy efficiency. Then, we design an automatic feature extraction scheme based on wavelet scattering and a CNN-LSTM neural network to achieve fine-grained gesture recognition. Finally, we model the handwashing gesture classification as a few-shot learning problem to mitigate the burden of collecting extensive data from five fingers. We collect data from 25 subjects over 2 months and evaluate the system performance on both commercial OURA ring and customized ring. Evaluation results show that WashRing achieves 97.8% accuracy which is 10.2%–15.9% higher than state-of-the-arts. Our adaptive sampling approach reduces energy consumption by 64.2% compared to fixed duty cycle sampling strategies.

Sports performance is often judged based on the results of a series of motions rather than observing and analyzing the detailed sequential motions that lead to the results. Hence, subjective feedback from the coaches is often ineffective in improving player performance. In this work, we custom-built a smart Internet of Things wristband motion sensor to implement data-based sports performance evaluation. A phase-based feature selection method is also proposed to assess the athletes’ sequential detailed motion for selected sport activities. To demonstrate the merits of this technology, we quantified the quality of the sequential motions of a specific type of volleyball serve by analyzing 183 samples of motion data obtained from a total of 18 players. The general skill levels (i.e., elite, subelite, and amateur) of the players were identified by machine learning algorithms, with accuracies of up to 95%. Moreover, we adopted biomechanical principles to extract 11 motion-related performance metrics from various phases of the players’ serve motion. We identified the distributions of these metrics across different skill levels and found eight key metrics that were highly correlated to the skill level of a players. We suggest that these metric distributions can be used as a reference for providing feedback to the coaches and players, to improve a player’s skill in the future. This phased-based analysis method can potentially be applied across many sports to increase the effectiveness of athletes’ training.

The aircrafts and ships in visible remote sensing (RS) images are of different scales. They are difficult to detect as they may be easily obscured by complex weather conditions such as snow and cloud. Therefore, it is important to eliminate the interference of complex weather conditions in order to detect these multi-scale objects accurately. This letter proposes an improved robust stacked multi-scale feature fusion network RSMNet to address this problem from two aspects. First, a stacked dilated convolution is used to enlarge the receptive fields of high-resolution images and improve the ability to extract multi-scale information. Second, the maps of extracted features are resized and integrated to refine the connection among different layers. Compared to the original Faster R-CNN model, RSMNet provides a 2% and 3.9% higher AP in the detection of aircrafts and ships, respectively. RSMNet also shows much more robust performance than the original model in detection under cloudy and snowy conditions.

Summary Impala system is an open source, analytic MPP database for Apache Hadoop. Impala system uses a query execution scheduling scheme that assigns near‐equal bytes retrieval tasks for different hosts to ensure system load balance. However, such “load balance” cannot guarantee a short response time for Impala system, when there are original loads in the system. Traditional query execution scheduling methods require either some assumptions or particular architecture, which cannot be directly used in Impala system. In this paper, we present a query execution scheduling scheme for Impala system. If the query fetches data from a single table, the scheme exploits the maximum flow algorithm. If the query fetches data from multiple tables, the scheme employs a cost‐based algorithm with heuristic pruning rules. In addition, we propose a cost model for Impala system, which considers parallel execution, communication cost, and cluster load. The performance of the proposed scheme is evaluated by the TPC‐DS benchmark, and experimental results show that the scheme can reduce the query response time by 10%‐30%.