Human Activity Recognition Method Based on FMCW Radar Sensor with Multi-Domain Feature Attention Fusion Network.

This paper proposes a human activity recognition (HAR) method for frequency-modulated continuous wave (FMCW) radar sensors. The method utilizes a multi-domain feature attention fusion network (MFAFN) model that addresses the limitation of relying on a single range or velocity feature to describe human activity. Specifically, the network fuses time-Doppler (TD) and time-range (TR) maps of human activities, resulting in a more comprehensive representation of the activities being performed. In the feature fusion phase, the multi-feature attention fusion module (MAFM) combines features of different depth levels by introducing a channel attention mechanism. Additionally, a multi-classification focus loss (MFL) function is applied to classify confusable samples. The experimental results demonstrate that the proposed method achieves 97.58% recognition accuracy on the dataset provided by the University of Glasgow, UK. Compared to existing HAR methods for the same dataset, the proposed method showed an improvement of about 0.9-5.5%, especially in the classification of confusable activities, showing an improvement of up to 18.33%.

Recognition of facial emotions across the lifespan: 8-year-olds resemble older adults.

On standard emotion recognition tasks with relatively long or unlimited stimuli durations, recognition improves as children grow older, whereas older adults are worse than young adults. Crucially, it was unknown (a) how older adults compare to age groups below young adulthood and (b) whether children can recognize emotions at shorter durations, with short durations likely common in real life. We compared emotion recognition in 5-year-olds, 8-year-olds, young adults and older adults at very brief durations (50 ms, 250 ms) as well as standard unlimited durations. Eight-year-olds were better than 5-year-olds, young adults than all groups, and there was a striking similarity between 8-year-olds and older adults, providing the first clear indication that older adults' recognition abilities are equivalent to that of an 8-year-old at all durations. Emotion recognition was above chance on all emotions and durations among the three older age groups and on most stimuli for 5-year-olds.

Cross Task Modality Alignment Network for Sketch Face Recognition.

The task of sketch face recognition refers to matching cross-modality facial images from sketch to photo, which is widely applied in the criminal investigation area. Existing works aim to bridge the cross-modality gap by inter-modality feature alignment approaches, however, the small sample problem has received much less attention, resulting in limited performance. In this paper, an effective Cross Task Modality Alignment Network (CTMAN) is proposed for sketch face recognition. To address the small sample problem, a meta learning training episode strategy is first introduced to mimic few-shot tasks. Based on the episode strategy, a two-stream network termed modality alignment embedding learning is used to capture more modality-specific and modality-sharable features, meanwhile, two cross task memory mechanisms are proposed to collect sufficient negative features to further improve the feature learning. Finally, a cross task modality alignment loss is proposed to capture modality-related information of cross task features for more effective training. Extensive experiments are conducted to validate the superiority of the CTMAN, which significantly outperforms state-of-the-art methods on the UoM-SGFSv2 set A, set B, CUFSF, and PRIP-VSGC dataset.

General cognitive decline does not account for older adults' worse emotion recognition and theory of mind.

Older adults have both worse general cognition and worse social cognition. A frequent suggestion is that worse social cognition is due to worse general cognition. However, previous studies have often provided contradictory evidence. The current study examined this issue with a more extensive battery of tasks for both forms of cognition. We gave 47 young and 40 older adults three tasks to assess general cognition (processing speed, working memory, fluid intelligence) and three tasks to assess their social cognition (emotion and theory-of-mind). Older adults did worse on all tasks and there were correlations between general and social cognition. Although working memory and fluid intelligence were unique predictors of performance on the Emotion Photos task and the Eyes task, Age Group was a unique predictor on all three social cognition tasks. Thus, there were relations between the two forms of cognition but older adults continued to do worse than young adults even after accounting for general cognition. We argue that this pattern of results is due to some overlap in brain areas mediating general and social cognition, but also independence, and with a differential rate of decline in brain areas dedicated to general cognition versus social cognition.

Face Manipulation Detection Based on Supervised Multi-Feature Fusion Attention Network.

Nowadays, faces in videos can be easily replaced with the development of deep learning, and these manipulated videos are realistic and cannot be distinguished by human eyes. Some people maliciously use the technology to attack others, especially celebrities and politicians, causing destructive social impacts. Therefore, it is imperative to design an accurate method for detecting face manipulation. However, most of the existing methods adopt single convolutional neural network as the feature extraction module, causing the extracted features to be inconsistent with the human visual mechanism. Moreover, the rich details and semantic information cannot be reflected with single feature, limiting the detection performance. Therefore, this paper tackles the above problems by proposing a novel face manipulation detection method based on a supervised multi-feature fusion attention network (SMFAN). Specifically, the capsule network is used for face manipulation detection, and the SMFAN is added to the original capsule network to extract details of the fake face image. Further, the focal loss is used to realize hard example mining. Finally, the experimental results on the public dataset FaceForensics++ show that the proposed method has better performance.

An Overdispersed Black-Box Variational Bayesian-Kalman Filter with Inaccurate Noise Second-Order Statistics.

Aimed at the problems in which the performance of filters derived from a hypothetical model will decline or the cost of time of the filters derived from a posterior model will increase when prior knowledge and second-order statistics of noise are uncertain, a new filter is proposed. In this paper, a Bayesian robust Kalman filter based on posterior noise statistics (KFPNS) is derived, and the recursive equations of this filter are very similar to that of the classical algorithm. Note that the posterior noise distributions are approximated by overdispersed black-box variational inference (O-BBVI). More precisely, we introduce an overdispersed distribution to push more probability density to the tails of variational distribution and incorporated the idea of importance sampling into two strategies of control variates and Rao-Blackwellization in order to reduce the variance of estimators. As a result, the convergence process will speed up. From the simulations, we can observe that the proposed filter has good performance for the model with uncertain noise. Moreover, we verify the proposed algorithm by using a practical multiple-input multiple-output (MIMO) radar system.

An Adaptive Ellipse Distance Density Peak Fuzzy Clustering Algorithm Based on the Multi-target Traffic Radar.

In the multi-target traffic radar scene, the clustering accuracy between vehicles with close driving distance is relatively low. In response to this problem, this paper proposes a new clustering algorithm, namely an adaptive ellipse distance density peak fuzzy (AEDDPF) clustering algorithm. Firstly, the Euclidean distance is replaced by adaptive ellipse distance, which can more accurately describe the structure of data obtained by radar measurement vehicles. Secondly, the adaptive exponential function curve is introduced in the decision graph of the fast density peak search algorithm to accurately select the density peak point, and the initialization of the AEDDPF algorithm is completed. Finally, the membership matrix and the clustering center are calculated through successive iterations to obtain the clustering result.The time complexity of the AEDDPF algorithm is analyzed. Compared with the density-based spatial clustering of applications with noise (DBSCAN), k-means, fuzzy c-means (FCM), Gustafson-Kessel (GK), and adaptive Euclidean distance density peak fuzzy (Euclid-ADDPF) algorithms, the AEDDPF algorithm has higher clustering accuracy for real measurement data sets in certain scenarios. The experimental results also prove that the proposed algorithm has a better clustering effect in some close-range vehicle scene applications. The generalization ability of the proposed AEDDPF algorithm applied to other types of data is also analyzed.

Use of a head camera to examine maternal input and its relation to 10- to 26-month-olds' acquisition of mental and non-mental state vocabulary.

We examined the relation between maternal responsiveness and children's acquisition of mental and non-mental state vocabulary in 59 pairs of mothers and children aged 10 to 26 months as they engaged in a free-play episode. Children wore a head camera and responsiveness was defined as maternal talk that commented on the child's actions (e.g., when the child reached for or manipulated an object visible in the head camera). As hypothesized, maternal responsiveness correlated with both mental and non-mental state vocabulary acquisition in younger children (approximately 18 months and younger) but not older children. We posit a diminishing role for maternal responsiveness in language acquisition as children grow older.

Music Making and Neuropsychological Aging: A Review.

Aging is associated with a decline in social understanding and general cognition. Both are integral to wellbeing and rely on similar brain regions. Thus, as the population ages, there is a growing need for knowledge on the types of activities that maintain brain health in older adulthood. Active engagement in music making might be one such activity because it places a demand on brain networks tapping into multisensory integration, learning, reward, and cognition. It has been hypothesized that this demand may promote plasticity in the frontal and temporal lobes by taxing cognitive abilities and, hence, increase resistance to age-related neurodegeneration. We examine research relevant to this hypothesis and note that there is a lack of intervention studies with a well-matched control condition and random assignment. Thus, we discuss potential causal mechanisms underlying training-related neuropsychological changes, and provide suggestions for future research. It is argued that although music training might be a valuable tool for supporting healthy neuropsychological aging and mental wellbeing, well-controlled intervention studies are necessary to provide clear evidence.