Semi-supervised object detection has made significant progress with the development of mean teacher driven self-training. Despite promising results, label mismatch problem is not yet fully explored in previous works, leading to severe confirmation bias during In this paper, we delve into and propose a simple effective LabelMatch framework from two different complementary perspectives, i.e., distribution-level instance-level. For former one, it reasonable approximate class distribution...

Abstract Infected wound repair stands as a formidable global issue, posing significant threats to public health and medical care. Biopolymer‐based dressings represent natural biocompatible candidate greatly conductive healing. However, the low productivity high price, coupled with widespread lack of antibacterial activities, vitally restrict their clinical applications. Herein, simple yet efficient top‐down assembly strategy is reported fabricate aggregation‐induced emission photosensitizer...

Traditional image classification often misclassifies unknown samples as known classes during testing, degrading recognition accuracy. Open-set can simultaneously detect (KCs) and (UCs) but still struggles to improve performance caused by open space risk. Therefore, we introduce a cosine distance loss function (CDLoss), which exploits the orthogonality of one-hot encoding vectors align with their corresponding encoder directions. This reduces overlap between feature spaces KCs UCs, mitigating...

Abstract This study investigates the fixed‐time bipartite consensus (FTBC) problem for leader–follower nonlinear multi‐agent systems (MASs) with nonzero control input leader and external disturbances. The cooperative–antagonistic relationships between agents are characterized by a structurally balanced signed graph. First, generalized Lipschitz‐type condition is introduced to realize of MASs. A distributed controller was then designed guarantee FTBC. upper bound convergence time sufficient...

In light of the escalating global concerns surrounding climate change, significance sustainable development in realm logistics cannot be overstated. This study undertakes imperative task devising strategies aimed at mitigating carbon emissions, reducing costs, minimizing transportation time, and enhancing customer satisfaction. The research delves into intricacies an optimization model tailored for a specific iteration Location-Routing Problem (LRP), namely Multi-Objective Multi-Period...

Abstract The accuracy of medium- and long-term runoff forecasting plays a significant role in several applications involving the management hydrological resources, such as power generation, water supply flood mitigation. Numerous studies that adopted combined models to enhance have been proposed. Nevertheless, some do not take into account effects different lag periods on selection input factors. Based this, this paper proposed novel model based periods. In approach, factors are initially...

Language learning has increasingly benefited from Computer-Assisted Learning (CALL) technologies, especially with Artificial Intelligence involved in recent years. CALL writing acknowledged as the core of language is being realized by technologies like Automated Writing Evaluation (AWE), and Essay Scoring (AES), which have developed considerably both computer education fields. AWE effectively enhanced EFL students’ performance to some extent, but such technology can only provide an...

In traditional financial analysis discriminant analyses with statistic methods are considered as primary and widely applied corporate credit rating bases in which human impact plays an essential role. Machine learning approaches enormously assist reducing interference improving efficiency. However, existing did not efficiently leverage the time series features of data that is beneficial for classification tasks. We propose a novel end-to-end architecture, SMAGRU, based on self multi-head...

The particle swarm optimization (PSO) is a wide used algorithm, which yet suffers from trapping in local optimum and the premature convergence. Many studies have proposed improvements to address drawbacks above. Most of them implemented single strategy for one problem or fixed neighborhood structure during whole search process. To further improve PSO performance, we introduced simple but effective method, named adaptive with Gaussian perturbation mutation (AGMPSO), consisting three...

Grammatical error correction has been considered as an application closely related to daily life and important shared task in many prestigious competitions workshops. The neural machine translation with encoder-decoder architecture containing language models the fundamental solution for grammatical correction. Whereas on texts of hearing impaired people or its not seen yet, common tasks are suffering several challenges, such insufficient training data, accuracy due unsatisfactory capacity...

Edge detection is important to image processing, and there are many approaches detect the edge features from image. However, high-resolution remotely sensed imagery including complex difficult extract features. This paper introduces a new method imagery. By using FFT, transformed into magnitude spectrum (frequency image). Then, frequency analyzed by Radius sampling angle sample. Finally, Gabor filter used setting proper parameter. The experimental result shows that proposed gives better...

Image matching is one of the most important problems in computer vision. Scale Invariant Feature Transform (SIFT) algorithm has been proved to be effective for detecting features image matching. However SIFT limitation extract textile or self-similar construction image. Fortunately fractional differentiation advantage strengthen and textural digital images. Aiming at problem, this paper proposes a new method based on SIFT. The calculates pyramid combining Riemann-Liouville (R-L) derivative...

Grounded language-image pre-trained models have shown strong zero-shot generalization to various downstream object detection tasks. Despite their promising performance, the rely heavily on laborious prompt engineering. Existing works typically address this problem by tuning text prompts using training data in a few-shot or fully supervised manner. However, rarely studied is optimize without any annotations. In paper, we delve into and propose an Unsupervised Prompt Tuning framework for...

The paper addresses the issue on landmark matching of images from Geosynchronous Earth Orbit (GEO) satellites. In general, satellite imagery is matched against base image, which predefined. When rotation occurs, accuracy many algorithms deteriorates. To overcome this problem, generalised Hough transform (GHT) employed for matching. At first an improved GHT algorithm proposed to enhance rotational invariance. Secondly a global coastline processed generate test image as and image. Then using...

Climate change, as an important environmental issue, has been widely investigated in recent decades. On the one hand, climate prediction is essential part for policy makers to response change of climate, which received many attentions. other there another challenging problem facing us today that some abnormal weathers occur globally, seems have relation e.g., global greenhouse effect, but with little existing researches on this relation. Therefore, paper, we propose two kinds climatic and...

Abstract In recent years, the application of neural networks in face recognition has improved its accuracy greatly. However, it still suffers from disappearance network gradient and training cost. this paper, we propose a new hybrid method which combines contrast divergence algorithm (CD) accelerated Restricted Boltzmann Machine (RBM) Integrated Learning Method (ILM) with Boosting (BM). First, use RBM to establish minimum energy model sample distribution. Then CD speed up feature extraction...

Cutset-type Possibilistic C-Means clustering (C-PCM) algorithm can significantly reduce the coincident phenomenon of (PCM) by introducing cut-set concept into PCM. The C-PCM also has strong robustness to noise and outliers. However, still suffers from center migration problem for datasets with small targets. In order solve this problem, a Semi-Supervised Possibility (SS-C-PCM) is proposed semi-supervised learning mechanism objective function utilizing some prior information guide process....

OOD-CV challenge is an out-of-distribution generalization task. To solve this problem in object detection track, we propose a simple yet effective Generalize-then-Adapt (G&A) framework, which composed of two-stage domain part and one-stage adaptation part. The implemented by Supervised Model Pretraining stage using source data for model warm-up Weakly Semi-Supervised both with box-level label auxiliary (ImageNet-1K) image-level performance boosting. as Source-Free Domain Adaptation paradigm,...

False positive is one of the most serious problems brought by agnostic domain shift in adaptive pedestrian detection. However, it impossible to label each box countless target domains. Therefore, yields our attention suppress false an unsupervised way. In this paper, we model object detection task into a ranking among and negative boxes innovatively, thus transform suppression problem re-ranking elegantly, which makes feasible solve without manual annotation. An attached during appears that...

At present, the safety management of construction mainly supervises similar incidents based on data accidents that have occurred, but ignores inherent uncertainty accidents. As a result, existing warning system is only suitable for single project and does not universality. In this paper, deep learning method YOLO used to detect violations such as wearing helmet smoking in industrial sites. algorithm improved by embedding convolutional block attention module (CBAM) an adaptive spatial feature...