Intelligent video surveillance plays a pivotal role in enhancing the infrastructure of smart urban environments. The seamless integration multi-angled cameras, functioning as perceptive sensors, significantly enhances pedestrian detection and augments security measures cities. Nevertheless, current pedestrian-focused target encounters challenges such slow speeds increased costs. To address these challenges, we introduce YOLOv5-MS model, an YOLOv5-based solution for detection. Initially,...

The problems of missed segmentation with fuzzy boundaries segmented regions and small are common in tasks, greatly decrease the accuracy clinicians' diagnosis. For this, a new network based on PSPNet, using coordinate attention mechanism mixed loss function for (CAML-PSPNet), is proposed. Firstly, module splits input feature map into horizontal vertical directions to locate edge position target. Then, Mixed Loss (MLF) introduced model training stage solve problem low small-target tumor...

In today's society, computer technology has been deeply rooted in the hearts of people. Computers are wonderful tools for creative thinking. It is an extension our visual function and cortex brain. Through this extension, we can see more scenes that could not before. As a simulation system creates feels virtual worlds, 3D digital reality uses as media or real imaginary scene. interactive dynamic vision entity behavior based on diversified information fusion. creator arts, must try to observe...

In this paper, a new QR code binarization method is presented according to the characteristic improve accuracy of decoding in mobile image processing. The Sauvola's thresholding algorithm modified solve which threshold decision uneven light condition. Experiments show that proposed has better correctness recognition compared with others.

In order to improve the accuracy of intelligent recognition marine ships besides realization algorithm, a deep learning technique is introduced applying Faster RCNN algorithm vessels. The target system structure based on established. effectiveness verified VOC2007 dataset, and vessel experiment carried out. On basis, database in different environments established marked, new dataset used train network model, trained model compared with old one by taking ship experiments. experimental results...

Large Language Models (LLMs) demonstrate impressive reasoning ability and the maintenance of world knowledge not only in natural language tasks, but also some vision-language tasks such as open-domain knowledge-based visual question answering (OK-VQA). As images are invisible to LLMs, researchers convert text engage LLMs into procedure. This leads discrepancies between their textual representations presented which consequently impedes final performance. To fill information gap better...

As part of a continuing research for evaluating threats posed exposed attack surface, this study will provide consolidated view exploitability vulnerable applications presenting web surface an organization to attacker. While testing and scanning technologies like Static Analysis Security Testing (SAST), Dynamic (DAST), Application Ethical Hack (Penetration Testing), monitoring technology the Web Firewall (WAF) provides traffic information number transaction requests every application under...

This research aims to build upon a conceptual idea of consolidating all application security vulnerability data from monitoring, detection, and discovery tools into physical system that allows for convergence observation response an event is threat. Multiple testing monitoring are deployed at different layers architecture capture activities occur layer. multi-layer disconnected without any analysis lineage the externally exposed web attack surface deep down layers. It only through this...

To deal with the serious visual artifacts caused by consistent restoration algorithm on local motion blurred images, a novel two stage blur region detection method including coarse location and refinement for images is proposed in this paper. First, feature discriminant criteria frequency domain spatial defined to generate image; then, binarization used image obtain map; subsequently, morphological methods are process clear areas trimap obtained. Finally, of achieved combination automatic...

In this paper, we address the text and image matching in cross-modal retrieval of fashion industry. Different from general domain, is required to pay much more attention fine-grained information images texts. Pioneer approaches detect region interests (i.e., RoIs) use RoI embeddings as representations. general, RoIs tend represent "object-level" images, while texts are prone describe detailed information, e.g. styles, attributes. thus not enough for matching. To end, propose FashionBERT,...

High-definition maps are crucial perception elements for autonomous robot navigation systems, which can provide accurate scene layout and environment information downstream motion prediction planning control tasks. Traditional methods based on manual annotation or SLAM algorithms require massive labor efforts time costs, hinders the deployment of practical applications. Online construction local from on-board cameras offers an alternative solution. Aiming at problems unsatisfying precision...

Abstract Accurate remaining useful life (RUL) prediction is vital for ensuring equipment safety, optimizing maintenance schedules, and reducing operational costs. Deep learning networks have emerged as effective means RUL prediction. However, these methods encounter challenges of multiple sensors data noise when handling large complex industrial equipment, which impede the model’s ability to generalize. To address issues, this paper proposes an improved Dynamical Variational Autoencoders...

A comprehensive equipment management system based on web service is introduced, and design idea described from the perspective of function scope, main functions integration.The key technologies foreground model background services are described.The integrated in integration capability extension ability, which can be manufacturing executi o n sy stem, i nformat ion executed rapidly.

To complement the insufficiency of real time and commonality in machine maintenance system, 95th Standard Project (SP95) standard was adopted for digitized plant modeling. Based on universal digital model, a configurable integrated system built with advanced technology real-time data collection mobile communication. Service-Oriented Architecture (SOA) Windows Communication Foundation (WCF) used to integrate ERP Inventory System, which provides fluent flow information activities. The key...