Local binary pattern (LBP) is a nonparametric descriptor, which efficiently summarizes the local structures of images. In recent years, it has aroused increasing interest in many areas image processing and computer vision shown its effectiveness number applications, particular for facial analysis, including tasks as diverse face detection, recognition, expression demographic classification. This paper presents comprehensive survey LBP methodology, several more variations. As typical...

A novel low-computation discriminative feature space is introduced for facial expression recognition capable of robust performance over a rang image resolutions. Our approach based on the simple local binary patterns (LBP) representing salient micro-patterns face images. Compared to Gabor wavelets, LBP features can be extracted faster in single scan through raw and lie lower dimensional space, whilst still retaining information efficiently. Template matching with weighted Chi square...

One essential problem in skeleton-based action recognition is how to extract discriminative features over all skeleton joints. However, the complexity of recent State-Of-The-Art (SOTA) models for this task tends be exceedingly sophisticated and over-parameterized. The low efficiency model training inference has increased validation costs architectures large-scale datasets. To address above issue, advanced separable convolutional layers are embedded into an early fused Multiple Input Branches...

RGB-induced salient object detection has recently witnessed substantial progress, which is attributed to the superior feature learning capability of deep convolutional neural networks (CNNs). However, such detections suffer from challenging scenarios characterized by cluttered backgrounds, low-light conditions and variations in illumination. Instead improving RGB based saliency detection, this paper takes advantage complementary benefits thermal infrared images. Specifically, we propose a...

Current methods for skeleton-based human action recognition usually work with complete skeletons. However, in real scenarios, it is inevitable to capture incomplete or noisy skeletons, which could significantly deteriorate the performance of current when some informative joints are occluded disturbed. To improve robustness models, a multi-stream graph convolutional network (GCN) proposed explore sufficient discriminative features spreading over all skeleton joints, so that distributed...

Particle filter and mean shift are two successful approaches taken in the pursuit of robust tracking. Both them have their respective strengths weaknesses. In this paper, we proposed a new tracking algorithm, embedded particle (MSEPF), to integrate advantages methods. Compared with conventional filter, MSEPF leads more efficient sampling by shifting samples neighboring modes, overcoming degeneracy problem, requires fewer particles maintain multiple hypotheses, resulting low computational...

This paper presents an overview of the state art three different fields with shared characteristics making use a network sensors, possible application computer vision, signal processing, and machine learning algorithms. Namely, first reports future directions for Intelligent Video Surveillance (IVS) applications, by recaping history field in terms hardware algorithmic progresses. Then, existing technologies Wireless Sensor Networks (WSNs) are compared described. Their applications to human...

Integration of multi-level contextual information, such as feature maps and side outputs, is crucial for Convolutional Neural Networks (CNNs)-based salient object detection. However, most existing methods either simply concatenate or calculate element-wise addition thus failing to take full advantages them. In this paper, we propose a new strategy guiding information integration, where outputs across layers are fully engaged. Specifically, shallower-level guided by the deeper-level learn...

Automated segmentation of brain glioma plays an active role in diagnosis decision, progression monitoring and surgery planning. Based on deep neural networks, previous studies have shown promising technologies for segmentation. However, these approaches lack powerful strategies to incorporate contextual information tumor cells their surrounding, which has been proven as a fundamental cue deal with local ambiguity. In this work, we propose novel approach named Context-Aware Network (CANet)...

Cerebrovascular diseases (CVDs) remain a leading cause of global disability and mortality. Digital Subtraction Angiography (DSA) sequences, recognized as the gold standard for diagnosing CVDs, can clearly visualize dynamic flow reveal pathological conditions within cerebrovasculature. Therefore, precise segmentation cerebral arteries (CAs) classification between their main trunks branches are crucial physicians to accurately quantify diseases. However, achieving accurate CA in DSA sequences...

Vision-based human affect analysis is an interesting and challenging problem, impacting important applications in many areas.In this paper, beyond facial expressions, we investigate affective body gesture video sequences, a relatively understudied problem.Spatial-temporal features are exploited for modeling of gestures.Moreover, present to fuse expression at the feature level using Canonical Correlation Analysis (CCA).By establishing relationship between two modalities, CCA derives semantic...

Local Binary Patterns (LBP) have been well exploited for facial image analysis recently. In the existing work, LBP histograms are extracted from local regions, and used as a whole regional description. However, not all bins in histogram necessary to be useful representation. this paper, we propose learn discriminative LBP-Histogram (LBPH) task of expression recognition. Our experiments illustrate that selected LBPH provide compact We experimentally it is consider multiscale representing...

<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Objective:</i> The utilization of hyperspectral imaging (HSI) in real-time tumor segmentation during a surgery have recently received much attention, but it remains very challenging task. xmlns:xlink="http://www.w3.org/1999/xlink">Methods:</i> In this work, we propose semantic methods, and compare them with other relevant deep learning algorithms for tongue segmentation. To the best our...

The primary treatment for malignant brain tumors is surgical resection. While gross total resection improves the prognosis, a supratotal may result in neurological deficits. On other hand, accurate intraoperative identification of tumor boundaries be very difficult, resulting subtotal resections. Histological examination biopsies can used repeatedly to help achieve but this not practically feasible due turn-around time tissue analysis. Therefore, techniques recognize types are investigated...

One essential problem in skeleton-based action recognition is how to extract discriminative features over all skeleton joints. However, the complexity of State-Of-The-Art (SOTA) models this task tends be exceedingly sophisticated and over-parameterized, where low efficiency model training inference has obstructed development field, especially for large-scale datasets. In work, we propose an efficient but strong baseline based on Graph Convolutional Network (GCN), three main improvements are...

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...

In this paper, we extensively investigate local features based facial expression recognition with face registration errors, which has never been addressed before. Our contributions are three fold. Firstly, propose and experimentally study the histogram of oriented gradients (HOG) descriptors for representation. Secondly, present representations on binary patterns (LBP) ternary (LTP) extracted from overlapping regions. Thirdly, quantitatively impact errors using different representations....

Zero-shot learning aims to recognize objects which do not appear in the training dataset. Previous prevalent mapping-based zero-shot methods suffer from projection domain shift problem due lack of image classes stage. In order alleviate problem, a deep unbiased embedding transfer (DUET) model is proposed this paper. The DUET composed (DET) module and an unseen visual feature generation (UVG) module. DET module, novel combined net integrates complementary merits linear nonlinear mapping...