Automatic emotion recognition based on multi-channel neurophysiological signals, as a challenging pattern task, is becoming an important computer-aided method for emotional disorder diagnoses in neurology and psychiatry. Traditional approaches require designing extracting range of features from single or multiple channel signals extensive domain knowledge. This may be obstacle non-domain experts. Moreover, traditional feature fusion can not fully utilize correlation information between...

In this paper, we propose a new research problem on active learning from data streams, where volumes grow continuously, and labeling all is considered expensive impractical. The objective to label small portion of stream which model derived predict future instances as accurately possible. To tackle the technical challenges raised by dynamic nature data, i.e., increasing evolving decision concepts, classifier-ensemble-based framework that selectively labels streams build classifier ensemble....

Computer vision systems today fail frequently. They also abruptly without warning or explanation. Alleviating the former has been primary focus of community. In this work, we hope to draw community's attention latter, which is arguably equally problematic for real applications. We promote two metrics evaluate failure prediction. show that a surprisingly straightforward and general approach, call ALERT, can predict likely accuracy (or failure) variety computer - semantic segmentation,...

Existing optical remote sensing image change detection (CD) methods aim to learn an appropriate discriminate decision by analyzing the feature information of bitemporal images obtained at same place. However, complex scenes in high-resolution (HR) cause unsatisfied results, especially for some irregular and occluded objects. Although recent self-attention-driven models with CNN achieve promising effects, computational consumed parameters costs emerge as impassable gap HR images. In this...

Feature selection is an important task in the analysis of hyperspectral data. Recently developed methods for learning sparse classifiers, which combine automatic feature and classifier design, established themselves among state art literature machine learning. In this letter, multinomial logistic regression (SMLR) introduced into community remote sensing utilized classification To relieve heavy degeneration performance caused by characteristics data oversparsity when SMLR selects a small...

In this paper, a real-time vehicle behavior analysis system is presented, which can be used in traffic jams and under complex weather conditions. recent years, many works based on background estimation foreground extraction for event detection have been reported. these studies, the images need to accurately segmented, although uneven illumination, shadows, overlapping are difficult handle. The main contribution of paper make point tracking without image segmentation procedure. proposed...

COVID-19 is an infectious disease caused by virus SARS-CoV-2 virus.Early classification of essential for cure and control.Transcription-polymerase chain reaction (RT-PCR) used widely the detection COVID-19.However, its high cost, time-consuming low sensitivity will significantly reduce diagnosis efficiency increase difficulty COVID-19.For X-ray images patients have inter-class similarity intra-class variability, we specifically designed a multi attention interaction enhancement module (MAIE)...

Analyzing wetland landscape pattern evolution is crucial for managing resources. High-resolution remote sensing serves as a primary method monitoring patterns. However, the complex types and spatial structures of wetlands pose challenges, including interclass similarity intraclass heterogeneity, leading to low separability landscapes difficulties in identifying fragmented small objects. To address these issues, this study proposes multilevel feature cross-fusion classification network...

Change detection methods for optical remote sensing images play an important role in environmental resource management. Although recent based on deep learning demonstrate incredible ability by constructing networks: 1)Extracting bi-temporal features a separate manner; 2)Fusing before forwarding them into the single-level network. Both severely neglect effect of spatial-temporal feature correlation between images. In addition, most existing represent multi-scale pairs layer-wise manner like...

An effective network structure is essential for the classification of satellite image time series (SITS). Deep learning models have been widely used SITS and achieved impressive performance, especially architectures based on self-attention. However, lack efficient comprehensive attention to valuable bands hinders performance some extent. To address this problem, an end-to-end attention-aware dynamic self-aggregation (ADSN) proposed in work, which combines two main parts: spectral focusing...

Insulin antibodies (IAs) affect blood glucose control in patients receiving insulin therapy.

Video super-resolution (VSR) remains challenging for real-world applications due to complex and unknown degradations. Existing methods lack the flexibility handle video sequences with different degradation levels, thus failing reflect scenarios. To address this problem, we propose a degradation-adaptive network (DAVSR) based on bidirectional propagation network. Specifically, adaptively employ three distinct levels process input sequences, aiming obtain training pairs that variety of...

Abstract Primary gastric adenosquamous carcinoma (PGASC) is a rare type of cancer with limited research and poorly understood clinicopathological features. This study investigated the features outcomes PGASC. Patients PGASC from Union Hospital, Tongji Medical College, Huazhong University Science Technology published literature were enrolled in this study. Survival curves generated using Kaplan–Meier method, prognostic factors identified through Cox proportional hazards regression models. 76...

Real-time processing of anomaly detection has become one the most important issues in hyperspectral remote sensing. Due to fact that widely used imaging spectrometers work a pushbroom fashion, it is necessary process incoming data line causal linewise progressive manner with no future involved. In this study, we proposed several processes well improve computational performance real-time (RCLP-AD). At first, Cholesky decomposition along linear system solving (CDLSS) was since background...

Hyperspectral image (HSI) classification is an important application of HSI analysis, which aims at assigning a class label to each pixel. However, considering that mixed pixels commonly exist in HSI, unique pixel imprecise. To better analysis the scene imaged we propose multi-label hyperspectral approach based on deep learning this study. First, stacked denoising autoencoder (SDAE) method used extract features for without supervision, can well represent nonlinearity high dimensional feature...