Many RGB-D visual attention models have been proposed with diverse fusion models; thus, the main challenge lies in differences results between different models. To address this challenge, we propose a local-global model for fixation prediction on an image; method combines global and local information through content-aware module (CAFM) structure. First, it comprises channel-based upsampling block exploiting contextual scaling up to same resolution as input. Second, our Deconv contains...

Fine-Grained Image Classification (FGIC) aims to classify images into specific subordinate classes of a superclass. Due insufficient training data and confusing samples, FGIC may produce uncertain classification results that are untrusted for applications. In fact, can be viewed as hierarchical process the multilayer information facilitates reduce uncertainty improve reliability FGIC. this paper, we adopt evidence theory measure confidence in propose trusted method through fusing evidence....

Lane keeping and lane change are the most general behaviors in highway driving. It's crucial to predict intention trajectory of surrounding vehicles accurately as it can help intelligent make safe decisions have a better understanding environment, thus improving safety autonomous driving promoting cooperative This paper focuses on prediction expressway scenarios. A dynamic bayesian network is modeled, based which probability each maneuver inferred. Then, founded estimation, roughly predicted...

Human eye-fixation prediction in 3D images is important for many applications, such as fine-grained video object segmentation and intelligent bulletproof curtains. While the vast majority of existing 2D-based approaches cannot be applied, main challenge lies inconsistency, or even conflict, between RGB depth saliency maps. In this paper, we propose a three-stream architecture to accurately predict human visual attention on end-to-end. First, two-stream feature extraction network based...

Freespace detection holds a crucial role in autonomous driving technology, particularly unstructured offroad scenarios that present additional challenges compared to structured road environments. Multimodal fusion methods are widely recognized as effective strategies for addressing these challenges. However, current often overlook the critical issue of noise interference multimodal data. To address this issue, we propose novel Noise-Aware Intermediary Fusion Network off-road freespace...

A large-scale and high-quality training dataset is an important guarantee to learn ideal classifier for text sentiment classification. However, manually constructing such a with labels labor-intensive time-consuming task. Therefore, based on the idea of effectively utilizing unlabeled samples, synthetical framework that covers whole process semi-supervised learning from seed selection, iterative modification set, co-training strategy proposed in this paper To provide basis selecting texts...

In this paper; the impedance model of PWM rectifier and LLC resonant converter are deduced, stability analysis cascade system is studied. The principle three-phase introduced; small signal in d-q coordinate deduced. expression dc side output for derived. operated a fixed-frequency state, modeled as model. On basis, input According to criterion, it can seen that amplitude greater than certain frequency domain. addition, Nyquist curve not around point (−1,0), which judge stable. simulation...

In order to reduce the false matching rate when detecting copy-move forgeries, an improved method based on SIFT and gray level was proposed in this study. Firstly, extract key points, establish feature vector for every point; Secondly, combine it with found a size of 129D; Finally, match above between two different points then regions would be detected. The experimental results showed that algorithm reduced even image distorted by Gaussian blur.

An HIV infection model with saturated rate and double delays is investigated. First, the existence of infection-free equilibrium [Formula: see text], immune-exhausted text] infected immunity in different conditions shown. By analyzing characteristic equation, we study locally asymptotical stability trivial equilibrium, Hopf bifurcations when two are used as bifurcation parameter. Furthermore, apply Nyquist criterion to estimate length delay for which continues hold. Then suitable Lyapunov...

Online dynamic security analysis conclusion is lagging power grid operation mode changes, and contradictions with the timeliness of real-time scheduling control have become increasingly prominent. To solve problem, framework online trend system constructed, three advanced technologies are researched, including future flow algorithm based on controlling multi transmission section simultaneously, method stability assessment index multiple fault coordination method. Huge developed, problem...

Dynamic security assessment (DSA) is widely used in dispatching operation systems, and the small signal stability one of DSA's most time-consuming calculation methods. In this article, a fast method proposed aiming to predict metrics designated oscillation mode, for example frequency or damping ratio. The much faster than simulation suitable online application. First, t -distributed stochastic neighbour embedding (t -SNE) algorithm performed which can create mapping from power system...

In terms of the accumulation historical data power system, China's grids have accumulated a great number dispatching and operating which contains different voltage levels. Online safety stability analysis conduct comprehensive system assessment every 15 minutes, including flow, short circuit, static analysis, stability, transient small disturbance limit transmission power, scheduling assistance decision-making, calculation result are about 1G. Based on six-month online by provincial grid,...

Bladder cancer is a malignant disease with substantial morbidity and mortality. staging crucial to determine the effective treatments of bladder tumors in clinic. As superiority feature learning, Deep Convolutional Neural Networks (DCNN) are widely used predict stage based on medical images. However, most existing DCNN-based methods data-driven neglect domain knowledge experiences clinicians. Besides, deep neural networks short model interpretability may lead risky diagnosis. To tackle...

Design DDS(Direct Digital Synthesizers, DDS) based on FPGA combine the advantage of output relatively bandwidth and frequency switching time. The design implementation digital synthesizer by are introduced in this article, moreover, simulated wave presented.

Domain adaptation aims to learn a classifier for target domain task by using related labeled data from the source domain. Because and may be mismatched, there is an uncertainty of with respect task. Ignoring lead models unreliable suboptimal classification results However, most previous works focus on reducing gap in distribution between domains. They do not consider about cannot apply adaptive classifier. Aimed at this problem, we revisit based evidence theory thereby devise measure. Based...

Domain adaptation aims to build up an adaptive model for the learning tasks on target domain by using data from source domain. Existing methods focused reducing gap between distributions of and but neglected uncertainty cross-domain learning. Actually, there exit in that are uncertain with respect task The will cause negative transfer impacts lead untrustworthy adaptation. Aiming at problem, we propose a measure based evidence theory represent classification utilize construct trusted deep...

<title>Abstract</title> Chlorpyrifos (CPF) is an insecticide and acaricide that interrupts nerve conduction by reducing the activity of enzyme acetylcholinesterase. It can enter human body through a variety channels produce serious physiological effects. However, excessive use CPF in herbal production produces pesticide residues reduces yield high-quality herbs. Therefore, development effective convenient residue detection technology critical for monitoring as well danger avoidance. Based on...