In recent years, growing needs for advanced security and traffic management have significantly heightened the prominence of visible-infrared person re-identification community (VI-ReID), garnering considerable attention. A critical challenge in VI-ReID is performance degradation attributable to label noise, an issue that becomes even more pronounced cross-modal scenarios due increased likelihood data confusion. While previous methods achieved notable successes, they often overlook...

Sufficient training data normally is required to train deeply learned models. However, due the expensive manual process for labelling large number of images, amount available always limited. To produce more a deep network, Generative Adversarial Network (GAN) can be used generate artificial sample data. generated usually does not have annotation labels. solve this problem, in paper, we propose virtual label called Multi-pseudo Regularized Label (MpRL) and assign it With MpRL, will as...

Cross-domain person re-identification (re-ID) is challenging due to the bias between training and testing domains. We observe that if backgrounds in datasets are very different, it dramatically introduces difficulties extract robust pedestrian features, thus compromises cross-domain re-ID performance. In this paper, we formulate such problems as a background shift problem. A Suppression of Background Shift Generative Adversarial Network (SBSGAN) proposed generate images with suppressed...

Current person re-identification (re-ID) works mainly focus on the short-term scenario where a is less likely to change clothes. However, in long-term re-ID scenario, has great chance A sophisticated system should take such changes into account. To facilitate study of re-ID, this paper introduces large-scale dataset called "Celeb-reID" community. Unlike previous datasets, same can clothes proposed Celeb-reID dataset. Images are acquired from Internet using street snap-shots celebrities....

Long-Term person re-identification (LT-reID) exposes extreme challenges because of the longer time gaps between two recording footages where a is likely to change clothing. There are types approaches for LT-reID: biometrics-based approach and data adaptation based approach. The former one seek clothing irrelevant biometric features. However, seeking high quality feature main concern. latter adopts fine-tuning strategy by using with significant change. performance compromised when it applied...

Accurate stride-length estimation is a fundamental component in numerous applications, such as pedestrian dead reckoning, gait analysis, and human activity recognition. The existing algorithms work relatively well cases of walking straight line at normal speed, but their error overgrows complex scenes. Inaccurate walking-distance leads to huge accumulative positioning errors reckoning. This paper proposes TapeLine, an adaptive algorithm that automatically estimates pedestrian’s using the...

In this paper, we propose an automatic engagement prediction method for the Engagement in Wild sub-challenge of EmotiW 2018. We first design a novel Gaze-AU-Pose (GAP) feature taking into account information gaze, action units and head pose subject. The GAP is then used subsequent level prediction. To efficiently predict long-time video, divide video multiple overlapped clips extract each clip. A deep model consisting Gated Recurrent Unit (GRU) layer fully connected as predictor. Finally,...

This paper considers person re-identification (re-ID) in the case of long-time gap (i.e., long-term re-ID) that concentrates on challenge clothes variation each person. We introduce a new dataset, named Celebrities-reID to handle challenge. Compared with current datasets, proposed dataset is featured two aspects. First, it contains 590 persons 10,842 images, and does not wear same clothing twice, making largest re-ID date. Second, comprehensive evaluation using state arts carried out verify...

The task of infrared-visible person re-identification (IV-reID) is to recognize people across two modalities ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i.e.</i> , RGB and IR). Existing cutting-edge approaches normally use a pair images that have the same IDs ID-tied cross-modality image pairs) input them into an ImageNet-trained ResNet50. ResNet50 backbone model can learn shared features tolerate modality discrepancies between IR. This...

Recent studies have seen significant advancements in the field of long-term person re-identification (LT-reID) through use clothing-irrelevant or insensitive features. This work takes a step further by addressing previously unexplored issue, Clothing Status Distribution Shift (CSDS). CSDS refers to differing ratios samples with clothing changes those without between training and test sets, leading decline LT-reID performance. We establish connection performance CSDS, argue that can improve...

Densely-sampled light field (LF) image is drawing increased attention for its wide applications in 3D reconstruction, digital refocusing, depth estimation, and virtual/augmented reality, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">et al.</i> In order to reconstruct a densely-sampled LF with high angular resolution, many computational methods have been proposed. However, most existing consider reconstruction based on neighboring views,...

Wi-Fi-based device-free human activity recognition has recently become a vital underpinning for various emerging applications, ranging from the Internet of Things (IoT) to Human–Computer Interaction (HCI). Although this technology been successfully demonstrated location-dependent sensing, it relies on sufficient data samples large-scale which is enormously labor-intensive and time-consuming. However, in real-world location-independent sensing crucial indispensable. Therefore, how alleviate...

A diffusion probabilistic model (DPM), which constructs a forward process by gradually adding noise to data points and learns the reverse denoising generate new samples, has been shown handle complex distribution. Despite its recent success in image synthesis, applying DPMs video generation is still challenging due high-dimensional spaces. Previous methods usually adopt standard process, where frames same clip are destroyed with independent noises, ignoring content redundancy temporal...

Food microbiological detection is an important part of food safety management. Understanding the behavior and characteristics microorganisms at cellular molecular level can enhance process. To quickly effectively detect microorganisms, a rapid method based on biotechnology computer vision proposed. Firstly, bacterial strains are cultivated sample data prepared. At level, this involves understanding growth kinetics metabolic processes microorganisms. Secondly, classification model proposed...

Person re-identification aims to match person captured by multiple non-overlapping cameras that mainly mean standard RGB cameras. In contemporary surveillance, of different modalities such as infrared and depth are introduced because their unique advantages in poor illumination scenarios. However, re-identifying the persons across is extremely difficult and, unfortunately, seldom discussed. It caused appearances shown under camera modalities. this paper, we tackle challenging cross-modality...

Device-free sensing (DFS) is an emerging technology that empowers wireless communication systems with the ability for not only data but also smart sensing. By taking advantage of machine-learning technologies, DFS transforms traditional networks into intelligent context-aware and will open doors a myriad promising 6G-enabled Internet Things (IoT) applications, ranging from home to buildings. Although significant progress has been made human activity recognition at single location by...

Synthetic high-resolution (HR) \& low-resolution (LR) pairs are widely used in existing super-resolution (SR) methods. To avoid the domain gap between synthetic and test images, most previous methods try to adaptively learn synthesizing (degrading) process via a deterministic model. However, some degradations real scenarios stochastic cannot be determined by content of image. These models may fail model random factors content-independent parts degradations, which will limit performance...

Seedling establishment is critical for grain yield and net benefits of wheat in wet clay soil after puddled rice harvest the Yangtze River basin (YRB) China. A wet-resistant rotary strip-till seeder (WR seeder) was developed to drill seeds zero tillage (ZT) conditions with complete residue mulching. The blades adopted were medium radius C type creating a 50–60 mm wide 30–50 deep furrow seed placement. pressing roller, usually used traditional seeders, has been replaced by two ground wheels...

Foliage penetration (FOPEN) has been found to be a critical mission for variety of applications, ranging from surveillance military. Recently, an emerging technology, namely wireless sensor network (WSN)-based device-free sensing (DFS), introduced the domain FOPEN. This technology only utilizes radio-frequency signals target detection and classification; thus, no additional hardware is required, just transceiver. Although feasibility using this human indoors explored some extent, it...