Object detection in aerial images is an active yet challenging task computer vision because of the bird’s-eye view perspective, highly complex backgrounds, and variant appearances objects. Especially when detecting densely packed objects images, methods relying on horizontal proposals for common object often introduce mismatches between Region Interests (RoIs) This leads to misalignment final classification confidence localization accuracy. In this paper, we propose a RoI Transformer address...

In this paper, we focus on tackling the problem of automatic accurate localization detected objects in high-resolution remote sensing images. The two major problems for object images caused by complex context information such contain are achieving generalizability features used to describe and locations. To address these challenges, propose a new framework, which can be divided into three processes: region proposal, classification, process. First, proposal method is generate candidate...

The past years have witnessed great progress on remote sensing (RS) image interpretation and its wide applications. With RS images becoming more accessible than ever before, there is an increasing demand for the automatic of these images. In this context, benchmark datasets serve as essential prerequisites developing testing intelligent algorithms. After reviewing existing in research community interpretation, article discusses problem how to efficiently prepare a suitable dataset...

Object detection in aerial images is an active yet challenging task computer vision because of the birdview perspective, highly complex backgrounds, and variant appearances objects. Especially when detecting densely packed objects images, methods relying on horizontal proposals for common object often introduce mismatches between Region Interests (RoIs) This leads to misalignment final classification confidence localization accuracy. Although rotated anchors have been used tackle this...

Automatically detecting airports from remote sensing images has attracted significant attention due to its importance in both military and civilian fields. However, the diversity of illumination intensities contextual information makes this task difficult. Moreover, auxiliary features within surrounding regions interest are usually ignored. To address these problems, we propose a novel method that uses multiscale fusion feature represent complementary each region proposal, which is extracted...

Because of recent advances in Convolutional Neural Networks (CNNs), traditional CNNs have been employed to extract thousands codes as feature representations for image retrieval. In this paper, we propose that more powerful features high-resolution remote sensing can be learned using only several tens codes; approach improve the retrieval accuracy and decrease time storage requirements. To accomplish goal, first investigate learning a series with different dimensions few via our improved CNN...

Abstract A major challenge in feature matching is the lack of objective criteria to determine corresponding points. Recent methods find match candidates first by exploring proximity descriptor space, and then rely on a ratio‐test strategy final correspondences. However, these measurements are heuristic subjectively excludes massive true positive correspondences that should be matched. In this paper, we propose novel algorithm for image collections, which capable providing quantitative...

Safe reinforcement learning (RL) studies problems where an intelligent agent has to not only maximize reward but also avoid exploring unsafe areas. In this study, we propose CUP, a novel policy optimization method based on Constrained Update Projection framework that enjoys rigorous safety guarantee. Central our CUP development is the newly proposed surrogate functions along with performance bound. Compared previous safe RL methods, benefits of 1) generalizes generalized advantage estimator...

Given an aerial image, scene parsing (ASP) targets to interpret the semantic structure of image content, e.g., by assigning a label every pixel image. With popularization data-driven methods, past decades have witnessed promising progress on ASP approaching problem with schemes tile-level classification or segmentation-based analysis, when using high-resolution images. However, former scheme often produces results tile-wise boundaries, while latter one needs handle complex modeling process...

Objects in remote sensing images are typically characterized with various appearances composed of complex spatial and spectral information, making the stable feature representation objects a difficult task. To address above issues, we propose new object detection method by combining CNN (Convolutional Neural Network) Swin Transformer. Specifically, first SCCA (Spatial-Channel Coordinate Attention) module to highlight essential features an image fusing spatial, channel, location information....

Object recognition, as one of the most fundamental and challenging problems in high-resolution remote sensing image interpretation, has received increasing attention recent years. However, conventional object recognition pipelines aim to recognize instances with bounding boxes a supervised learning strategy, which require intensive manual labor for instance annotation creation. In this paper, we propose weakly method alleviate problem. The core idea our is multiple objects an using only...

Incorporating contrastive learning objectives in sentence representation (SRL) has yielded significant improvements on many sentence-level NLP tasks. However, it is not well understood why works for semantics. In this paper, we aim to help guide future designs of methods by taking a closer look at SRL through the lens isotropy, contextualization and dynamics. We interpret its successes geometry shifts show that brings drives high intra-sentence similarity: when same sentence, tokens converge...

In the field of remote surveillance, acquiring high-quality voice target has always been an exciting goal. this paper, we propose a method based on convolutional neural network to extract target's speech signals remotely. The consists two parts: optical setup enables us obtain speckle images conveniently and covertly, model is used recover from continuous images. Correlation coefficient root mean square error metrics show effectiveness our for extraction. Compared traditional spatial image...

Manual segmentation of atherosclerotic plaque for quantitative assessment is a time-consuming process. In this study, convolutional neural network based automatic method named Vessel-Segnet was proposed evaluation lumen, vessel wall and on MR images. The achieved the best performance with highest dice similarity coefficient lowest average surface distance among six models. terms morphological evaluation, excellent agreement manual method. Overall, can quickly accurately realize evaluation.

Motivation: Manual centerline Extraction based on black-blood Magnetic resonance vessel wall imaging is a difficult and time-consuming but important step for further analysis of plaques. Goal(s): To propose method quickly, automatically accurately extracting the label segments target arteries. Approach: This study proposes that combines deep learning traditional graphics automatic accurate extraction even labeling, which applicable to flexible MR sequences (only images, only bright-blood or...

Motivation: With the assistance of prior vessel wall mask, segmentation atherosclerotic plaque can achieve satisfactory performance. However, manual sketching mask is still time-consuming. Goal(s): To propose a method for fast and accurate without relying on knowledge walls. Approach: This study proposes deep learning model based multi-head loss design automatic carotid artery plaques, with aim reducing dependence information walls in segmentation. Results: In independent test, achieving...

We study a typical optimization model where the variable is composed of multiple probability distributions. Though appears frequently in practice, such as for policy problems, it lacks specific analysis general setting. For this problem, we propose new structural condition/landscape description named generalized quasar-convexity (GQC) beyond realms convexity. In contrast to original \citep{hinder2020near}, GQC allows an individual quasar-convex parameter $\gamma_i$ each block $i$ and smaller...

Precipitation unevenness significantly influences the rational allocation of water resources and management agricultural irrigation. Based on precipitation data from 29 meteorological stations in Heilongjiang Province, China, 1961 to 2020, this study calculated concentration index (PCI), degree (PCD), period (PCP) analyze spatial distribution characteristics heterogeneity at three distinct timescales: year, maize growth period, four stages period. The findings reveal that rainy season...

Abstract A convenient and efficient method for one pot synthesis of polysubstituted ( Z )‐halobenzo[ c,d ]indoles from 8‐alkynyl‐1‐naphthylamine derivatives using CuCl 2 CuBr as the halogenated reagents has been developed. In this protocol, both (Z)‐chloro bromobenzo[ ] indole can be obtained in moderate to good yields by copper halide‐promoted stereoselective intramolecular cis ‐addition annulation reactions. novel fluorescent molecule with AIE properties was constructed via Suzuki coupling...

With the rapid advancement of remote sensing (RS) technology, RS image interpretation has made great progress and been widely used in broad applications, which constructed benchmark datasets for developing testing intelligent algorithms have playing an increasingly critical role. Motivated by essential prerequisites dataset development algorithms, this manuscript provides a discussion on construction interpretation. Specifically, we first analyze current challenges review widespread is...