The contents of working memory must be maintained in the face distraction, but updated when appropriate. To manage these competing demands stability and flexibility, representations are complemented by distinct gating mechanisms that selectively transmit information into out stores. operations such dopamine-dependent systems midbrain striatum their complementary maintenance cortex may therefore dissociable. If true, selective increases cortical dopamine tone should preferentially enhance...

Summary This article studies reachable set estimation for linear discrete‐time systems with time delay, which are influenced by unknown but bounded disturbances. We propose a novel method based on zonotopes the considered systems. The proposed can estimate real‐time under nonzero initial conditions. In order to increase accuracy, we an iterative reduce conservatism caused couplings between sets at different instants. effectiveness of is illustrated three numerical simulations.

In recent years, with the continuous development of multi-agent technology represented by unmanned aerial vehicle (UAV) swarm, consensus control has become a hot spot in academic research. this paper, we put forward discrete-time protocol and obtain necessary sufficient conditions for second-order system fixed structure under condition no saturation input. The theoretical derivation verifies that two eigenvalues Laplacian communication network matrix sampling period have an important effect...

Stitching images with parallax for naturalness remains a challenging problem. This paper proposes an image stitching method which preserves the flatness of planes in scene natural look. Our formulates alignment as camera parameters and normal vectors planes. Given set feature point matches, process grouping points into different layers rejecting outliers is introduced. According to epipolar constraint corresponding two images, focal length pose change are recovered simultaneously. Then,...

This article deals with the template matching problem, and a weighted smallest deformation similarity measure, which is robust to occlusions, background outliers, complex deformations. The appearance-based nearest neighbor (NN) of points constructed location distance between each point in its employed penalize explicitly. Then, weights are added relied on their likelihood belonging through NN around target window. Experiments show that proposed method improves state-of-the-art performance...

To resolve the issues of a deep backbone network, large model, slow reasoning speed on mobile terminal, low detection accuracy for small targets and difficulties detecting recognizing traffic lights in real time accurately with YOLOv4, recognition method based improved YOLOv4 is proposed. The lightweight ShuffleNetv2 network utilized to replace CSPDarkNet53 satisfy requirements terminal. reformed k-means clustering algorithm applied generate anchor boxes avoiding sensitivity issue outliers...

Homography estimation is a fundamental task in computer vision that involves obtaining the transformation between multi-view images for image alignment. Although convolutional neural networks (CNNs) have shown state-of-the-art performance this task, few works explored use of transformer-based models, which demonstrated superiority high-level tasks. In paper, we propose strong baseline model homography combines swin transformer feature representation global features and CNN local features....

Template matching in unconstrained environment with complex deformation, occlusion, and background clutter is a challenging task. Recently, some measures which are robust to outliers were presented, however, they fix the window size thus cannot handle large-scale change. In this article, multiscale template method based on nearest neighbor (NN) search proposed. To discover effect of scale measure, expectation diversity similarity (DIS) derived by probabilistic analysis. Then, scale-adaptive...

The couplings among different networks facilitate their communications, while at the same time they also bring risk of enhancing wide spread cascading failures to coupled networks. Given that there is usually time-delay during and more than one coupling link a node might possess, failure model for scale-free multi-coupling-link built in this paper, based on map lattices (CML) model, which may be wider representative previous models. Our research shows BA (Barabási-Albert) networks, threshold...

Nanoscale coating manufacturing (NCM) process modeling is an important way to monitor and modulate quality. The multivariable prediction of coated film the data augmentation NCM are two common issues in smart factories. However, there has not been artificial intelligence model solve these problems simultaneously. Focusing on problems, a novel auxiliary regression using self-attention-augmented generative adversarial network (AR-SAGAN) proposed this paper. This deals with problem three steps....

Self-localization is an important topic for mobile robot. Although some ceiling-based visual odometry have been proposed, the localization process still cumbersome. In this paper, we propose a novel monocular based on ceiling vision that fast and easy to implement. By pointing camera ceiling, reduces computation eliminates interference from dynamic environments. We apply Oriented FAST Rotated BRIEF (ORB) match features between two frames of view use sliding window filter remove incorrect...

Summary This article studies reachable set estimation for parameter uncertain systems with time delay. A new method is proposed nonlinear discrete‐time uncertainty and unknown but bounded disturbance. By analyzing Lyapunov‐Krasovskii functional, formulated to an optimization problem in terms of linear matrix inequalities. The not only can handle nonzero initial condition also provide real‐time estimation. effectiveness superiority the are illustrated by numerical simulations.

Logistic regression is a supervised binary classification algorithm in machine learning. It an important part of neural network and convolutional network. An the logistic to find optimal parameters loss function, which often non-linear convex optimization problem. Generally, gradient descent method or stochastic linear search are used this nonlinear problem, but these methods easy fall into trap local minimum have weak global convergence. In order preserve information better enhance...

ABSTRACT The water reuse facilities of industrial parks face the challenge managing a growing variety wastewater sources as their inlet water. Typically, this clustering outcome is designed by engineers with extensive expertise. This paper presents an innovative application unsupervised learning methods to classify in Chinese stations, aiming reduce reliance on engineer experience. concept ‘water quality distance’ was incorporated into three algorithms (K-means, DBSCAN, and AGNES), which...

Deep learning has recently demonstrated its excellent performance on the task of multi-view stereo (MVS). However, loss functions applied for deep MVS are rarely studied. In this paper, we first analyze existing functions' properties depth based approaches. Regression leads to inaccurate continuous results by computing mathematical expectation, while classification outputs discretized values. To end, then propose a novel function, named adaptive Wasserstein loss, which is able narrow down...