Depression constitutes a pressing global health burden, especially among young adults, a demographic that remains under-represented in computational mental health research. This paper presents a novel end-to-end PAVNet that detects depression in young adults, based on Personality-aware AudioVisual multimodal modeling. Our framework effectively integrates audio-visual cues, personality traits and longitudinal userlevel data through hierarchical fusion and adaptive soft gating mechanisms. Enhanced by automated clustering, the adaptive personality encoder dynamically models individual differences, while the user-level feature generation module aggregates multimodal information from multiple historical interactions to construct comprehensive user profiles. Extensive experiments on the MPDD-Young dataset establish state-of-the-art performance with an accuracy of 0.9564 and F1-score of 0.9458, demonstrating that multimodal fusion and personalized modeling markedly enhances both accuracy and robustness. These findings highlight the importance of integrating personality and longitudinal user data for scalable, personalized assessment for early detection and intervention in youth mental health. The key code is available at https://github.com/yilin-succeed/PAVNet.

Occlusion has always been a difficulty in multi-pedestrian detection and tracking. The foreground object occlusion and the overlapping self-occlusion bring many problems of false detection and identity loss. In this paper, we propose an adaptive attention-aware method for occluding pedestrian tracking, which belongs to detection-based online tracking. To deal with the person re-identification under short-term occlusion, we define a global-local occlusion coefficient and segment the occluded part of pedestrians by spatial-temporal attention module, so that the discriminative features can be adaptively extracted. During data association, the trajectory matching cost matrix is set to adjust the target pedestrian matching. Experimental results demonstrate the superior performance of our method in videos with severe occlusion, indicating its effectiveness for tracking occluded pedestrians.

Abstract In the field of multi-modal face anti-spoofing (MFAS), where RGB, depth, and infrared data are integrated, remarkable advancements have been seen. However, despite the advancement, there still exist challenges when it comes to adaptability, particularly in dealing with unseen attacks. In this paper, a novel model called MuST-GAN MFAS is presented. This model employs a generative network that incorporates modality-specific encoders and transformer layers. It is significant that the model efficiently disentangles multi-semantic spoof traces by utilizing the power of cross-modal attention mechanisms and a transformer-based spoof trace generator. The training process involves bidirectional adversarial learning, ensuring identity consistency, intensity, center, and classification losses are taken into consideration. Through precise evaluations, it has been shown that the proposed model surpasses existing frameworks, showing remarkable performance when evaluating several modal samples. In the end, MuST-GAN MFAS makes an impressive contribution to the field of face anti-spoofing by offering results that are easy to interpret and emphasizing how important it is to learn multi-semantic spoof traces in order to improve generalization and adaptability to unseen attacks. The code is available at https://github.com/ZainUlAbideenMalik/Must-GAN-MFAS.

Abstract Facial attractiveness prediction is an important research topic in the computer vision community. It not only contributes to the development of interdisciplinary research in psychology and sociology, but also provides fundamental technical support for applications like aesthetic medicine and social media. With the advances in 3D data acquisition and feature representation, this paper aims to investigate the facial attractiveness from deep learning and three‐dimensional perspectives. The 3D faces are first processed to unwrap the texture images and refine the raw meshes. The feature extraction networks for texture, point cloud, and mesh are then delicately designed, considering the characteristics of different types of data. A more discriminative face representation is derived by feature fusion for the final attractiveness prediction. During network training, the cyclical learning rate with an improved range test is introduced, so as to alleviate the difficulty in hyperparameter setting. Extensive experiments are conducted on a 3D FAP benchmark, where the results demonstrate the significance of deep feature fusion and enhanced learning rate in cooperatively facilitating the performance. Specifically, the fusion of texture image and point cloud achieves the best overall prediction, with PC, MAE, and RMSE of 0.7908, 0.4153, and 0.5231, respectively.

The multivariate fusion transformation is ubiquitous in multivariate time series prediction (MTSP) problems. The previous multivariate fusion transformation fuses the feature of different variates at a time step, then projects them to a new feature space for effective feature representation. However, temporal dependency is the most fundamental property of time series. The previous manner fails to capture the temporal dependency of the feature, which is destroyed in the transformed feature matrix. Multivariate feature extraction based on the feature matrix with missing temporal dependency leads to the loss of predictive performance of MTSP. To address this problem, we propose the Temporal Dependency Priority for Multivariate Time Series Prediction (TemDep) method. Specifically, TemDep extracts feature temporal dependency of multivariate time series first and then considers multivariate feature fusion. Moreover, the low-dimensional and high-dimensional feature fusion manners are designed with the temporal dependency priority to fit different dimensional multivariate time series. The extensive experimental results of different datasets show that our proposed method can outperform all state-of-the-art baseline methods. It proves the significance of temporal dependency priority for MTSP.

Facial expression recognition (FER) plays a significant role in our daily life. However, annotation ambiguity in the datasets could greatly hinder the performance. In this paper, we address FER task via label distribution learning paradigm, and develop a dual-branch Adaptive Distribution Fusion (AdaDF) framework. One auxiliary branch is constructed to obtain the label distributions of samples. The class distributions of emotions are then computed through the label distributions of each emotion to exclude ambiguity existing in distributions. Finally, those two distributions are adaptively fused according to the attention weights to train the target branch. Extensive experiments are conducted on three real-world datasets, RAF-DB, AffectNet and SFEW, where our Ada-DF shows advantages over the state-of-the-art works. The code is available at https://github.com/taylor-xy0827/Ada-DF.

Aspect sentiment triplet extraction (ASTE) is a challenging subtask in aspect-based sentiment analysis.It aims to explore the triplets of aspects, opinions and sentiments with complex correspondence from the context.The bidirectional machine reading comprehension (BMRC) can effectively deal with ASTE task, but several problems remains, such as query conflict and probability unilateral decrease.Therefore, this paper presents a robustly optimized BMRC method by incorporating four improvements.The word segmentation is applied to facilitate the semantic learning.Exclusive classifiers are designed to avoid the interference between different queries.A span matching rule is proposed to select the aspects and opinions that better represent the expectations of the model.The probability generation strategy is also introduced to obtain the predicted probability for aspects, opinions and aspect-opinion pairs.We have conducted extensive experiments on multiple benchmark datasets, where our model achieves the stateof-the-art performance.1

Diagnosis of renal cell carcinoma (RCC) is critical in automated clinical decision-support system. Existing state-of-the-art methods focus on designing complex machine learning models for high identification accuracy; especially, deep neural networks improve the prediction accuracy. Such designs ignore the explainability of models, and their “black box” nature is a barrier to model trust. In addition, little attention has been paid to evaluating clinical utility. To explain model predictions and evaluate risks and benefits, this paper introduces the explainable machine learning predictions that incorporate the balancing of risks and benefits of treatment to RCC prediction models. The proposed explainable network is based on tree ensembles with four improvements: (1) A multiscale feature extraction module, obtaining comprehensive radiomic features; (2) An attribute optimization module based on Chi-square test, guiding the network to focus on useful information at variables; (3) Appending a SHapley Additive exPlanations (SHAP) module to the framework to automatically and efficiently interpret the prediction of the models; And (4) a decision curve analysis (DCA) module is performed for the clinical utility evaluation. By integrating the above improvements in series, the models' performances are gradually enhanced. By the comparison of different tree ensembles-based algorithms, our study finds the random forest (RF) and extra trees (ET) classifier can be valuable diagnosis tools for explainable RCC predictions. To demonstrate the generalizability, our tree ensembles-based models achieve higher accuracies than the state-of-the-art pretrained deep models with fine-tuned parameters.

Anatomical landmark detection is the first step towards the analysis of 3D medical data. In this paper, we annotate the biologically significant landmarks on sphere-like meshes using deep semantic segmentation. A triplet candidate pool and the cutting path are firstly defined to parameterize 3D mesh model into 2D planar flat-torus. A deep convolutional network is utilized to learn geometric surface properties and then segment landmark areas. The landmarks are finally localized within their areas by incorporating the local neighborhood features. Extensive experiments are conducted on our newly-constructed scapula dataset, where we demonstrate the accuracy and efficacy of the proposed approach.

Two key challenges lie in the facial attractiveness computation research: the lack of discriminative face representations, and the scarcity of sufficient and complete training data. Motivated by recent promising work in face recognition using deep neural networks to learn effective features, the first challenge is expected to be addressed from a deep learning point of view. A very deep residual network is utilized to enable automatic learning of hierarchical aesthetics representation. The inspiration to deal with the second challenge comes from the natural representation of the training data, where each training face can be associated with a label (score) distribution given by human raters rather than a single label (average score). This paper, therefore, recasts facial attractiveness computation as a label distribution learning problem. Integrating these two ideas, an end-to-end attractiveness learning framework is established. We also perform feature-level fusion by incorporating the low-level geometric features to further improve the computational performance. Extensive experiments are conducted on a standard benchmark, the SCUT-FBP dataset, where our approach shows significant advantages over the other state-of-the-art work.

Facial attractiveness computation is a challenging task because of the lack of labeled data and discriminative features. In this paper, an end-to-end label distribution learning (LDL) framework with deep convolutional neural network (CNN) and geometric features is proposed to meet these two challenges. Different from the previous work, we recast this task as an LDL problem. Compared with the single label regression, the LDL could improve the generalization ability of our model significantly. In addition, we propose some kinds of geometric features as well as an incremental feature selection method, which could select hundred-dimensional discriminative geometric features from an exhaustive pool of raw features. More importantly, we find these selected geometric features are complementary to CNN features. Extensive experiments are carried out on the SCUT-FBP dataset, where our approach achieves superior performance in comparison to the state-of-the-arts.