检索结果-内蒙古大学图书馆

Repeatable adaptive keypoint detection via self-supervised learning

Science China(information Sciences) 2022年第11期65卷 123-147页

作者： Pei YAN Yihua TAN Yuan TAI National Key Laboratory of Science & Technology on Multi-spectral Information Processing School of Artificial Intelligence and Automation Huazhong University of Science and Technology

keypoint-based matching is a fundamental technology for different computer vision tasks, in which keypoint detection is a crucial step and directly affects the entire performance. Based on deep learning approaches, the learning-based keypoint detectors have been significantly developed. To further improve the accuracy of high-level matching tasks, the extracted keypoints should provide more accurate point-to-point correspondences and maintain a uniform spatial distribution. Based on this idea, a self-supervised learning method of keypoint detection named repeatable adaptive point is proposed. This method consists of a selfsupervised objective and an optimization algorithm. The objective maximizes the repeatability measure with the sparsity constraint of keypoints. This sparsity constraint is formulated by combining the non-maximum suppression operation and the penalty function of keypoint number, which generally makes keypoints have a uniform spatial distribution. A novel approximate alternate optimization algorithm is proposed to maximize the above objective, whose convergence is proved in theory. The proposed detector is “adaptive” because the combinations of it and some existing descriptors can adapt to high-level matching tasks with a fast convergence speed. Specifically, the combinations of it and SuperPoint/HardNet descriptors achieve stateof-the-art accuracy on three high-level tasks based on image matching, namely homography estimation,camera pose estimation, and three-dimensional reconstruction. Furthermore, the proposed method converges faster on new scenes compared with the state-of-the-art method that jointly optimizes the detector and the descriptor.

关键词： keypoint detection convolutional neural network image matching self-supervised learning repeatable

来源：评论

学校读者我要写书评

暂无评论

Distribution Mining and Noise Correction for Facial Expression Recognition

Distribution Mining and Noise Correction for Facial Expressi...

引用

Computer Vision, Image and Deep Learning (CVIDL), International Conference on

作者： Hailan Huang Yihua Tan National Key Laboratory of Science & Technology on Multi-spectral Information Processing School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China

ISBN: (数字)9798350373820

ISBN: (纸本)9798350373837

Due to the high annotation cost and inter-class ambiguity, many facial expression datasets only offer noisy one-hot labels. This creates two challenges for facial expression recognition (FER) as the one-hot labels only represent limited single semantics, and the noisy labels will mislead the model. Our study proposes a novel method to perform both label Distribution Mining and Noise Correction (DMNC) on the FER dataset. That is, while eliminating the disturbance of noisy samples, we start from the original one-hot labels and gradually refine them to distributions of expressions close to the real world. Specifically, DMNC uses the network forgetting property to select noisy samples. Then based on clean samples, an expression feature space is learned by contrastive learning. Finally, the labels of clean and noisy samples are refined or relabelled in a divide-and-conquer manner using the learned feature space. Experiments show that our method exhibits leading performance on both real-world FER benchmarks and synthetic noisy datasets.

关键词： Deep learning Computer vision Image recognition Costs Face recognition Semantics Noise

来源：评论

学校读者我要写书评

暂无评论

Non-Reference Quality Assessment Method Based on Occlusion Region information Estimation and Natural Scene Statistics

Non-Reference Quality Assessment Method Based on Occlusion R...

引用

2023 International Conference on Optics and Machine Vision, ICOMV 2023

作者： Lv, SiQi Ren, LiJun Zhang, YiQiang Huang, ChengYang OuYang, Yan He, Jia Hu, Jing Huazhong Univ. of Science and Technology Wuhan China Air Force Early Warning Academy Wuhan China Chinese People’s Liberation Army troops 95841 China National Key Laboratory of Science and Technology on Multi-Spectral Information Processing Wuhan China

ISBN: (纸本)9781510664777

Most of the laser interfered image quality assessment algorithms need to know the reference images or partial information of reference images. However, in practical application, the reference image or its related informa- tion is difficult to obtain, which makes the application scenario of laser interference image quality evaluation algorithm is greatly limited. To solve this problem, this paper starts with the prediction processing of the ob- scured information, and improves the Markov random field estimation algorithm (MRF) to realize the real-time estimation of the obscured area information. Then, proposes a non-reference image quality assessment method based on occlusion area information estimation and natural scene statistics(IENSS), which analyzes the statistical characteristics of laser interfered images in natural scenes. The model is trained by machine learning. Finally, simulation experiments are carried out to verify the effectiveness of the proposed method. © 2023 SPIE. All rights reserved.

关键词： Markov processes

来源：评论

学校读者我要写书评

暂无评论

Learning Compact and Overlap-Biased Interactions for Point Cloud Registration

SSRN

引用

SSRN 2024年

作者： Guo, Lin Chen, Zhi Cheng, Senmao Yang, Fan Tao, Wenbing National Key Laboratory of Science and Technology on Multi-spectral Information Processing School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China

Point cloud registration is a fundamental task in computer vision. Recent Transformer based methods for point cloud registration take advantage of the interaction modeling ability of the attention operation. However, as two of the main challenges of point cloud registration, feature ambiguity, and low overlap are still the bottleneck of the performance in real scenes. In this paper, we present a new neural network to solve these two problems in Transformer architecture. First, we analyze that the dense interaction of the cross-attention (CA) component is a factor for the feature ambiguity problem. Thus, we propose an Optimal Transport guided Cross Attention (OT-CA) to build compact interactions in CA. It uses a Spatial Consistency guided cost Regularization (SCR) to build the cost for the optimal transport problem, and get the weight matrix of CA by solving it. Through the structure information introduced by SCR and the more reasonable interactions in CA, the network can alleviate the feature ambiguity problem with fewer computing resources. Meanwhile, to solve the low-overlap problem, we propose a Separate-and-Joint Overlap Prediction module. It adopts separate branches and training steps for feature matching and overlap prediction to reduce negative impacts between these two tasks, and adopts a joint training process to make full use of overlap information for learning better feature matching. Finally, the proposed modules are embedded into a coarse-to-fine pipeline. Our method shows state-of-the-art performance on three benchmark datasets. © 2024, The Authors. All rights reserved.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

End-to-end Video Gaze Estimation via Capturing Head-face-eye Spatial-temporal Interaction Context

arXiv

引用

arXiv 2023年

作者： Guan, Yiran Chen, Zhuoguang Zeng, Wenzheng Cao, Zhiguo Xiao, Yang National Key Laboratory of Science and Technology on Multi-Spectral Information Processing School of Artificial Intelligence and Automation Huazhong University of Science and Technology China

In this letter, we propose a new method, multi-Clue Gaze (MCGaze), to facilitate video gaze estimation via capturing spatial-temporal interaction context among head, face, and eye in an end-to-end learning way, which has not been well concerned yet. The main advantage of MCGaze is that the tasks of clue localization of head, face, and eye can be solved jointly for gaze estimation in a one-step way, with joint optimization to seek optimal performance. During this, spatial-temporal context exchange happens among the clues on the head, face, and eye. Accordingly, the final gazes obtained by fusing features from various queries can be aware of global clues from heads and faces, and local clues from eyes simultaneously, which essentially leverages performance. Meanwhile, the one-step running way also ensures high running efficiency. Experiments on the challenging Gaze360 dataset verify the superiority of our proposition. The source code will be released at https://***/zgchen33/MCGaze. Copyright © 2023, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

SC$^{2}$2-PCR++: Rethinking the Generation and Selection for Efficient and Robust Point Cloud Registration

引用

IEEE Transactions on Pattern Analysis and Machine Intelligence 2023年第10期45卷 12358-12376页

作者： Zhi Chen Kun Sun Fan Yang Lin Guo Wenbing Tao National Key Laboratory of Science and Technology on Multi-spectral Information Processing School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan Hubei China Hubei Key Laboratory of Intelligent Geo-Information Processing School of Computer Science China University of Geosciences Wuhan Hubei China

Outlier removal is a critical part of feature-based point cloud registration. In this article, we revisit the model generation and selection of the classic RANSAC approach for fast and robust point cloud registration. For the model generation, we propose a second-order spatial compatibility (SC $^{2}$ ) measure to compute the similarity between correspondences. It takes into account global compatibility instead of local consistency, allowing for more distinctive clustering between inliers and outliers at an early stage. The proposed measure promises to find a certain number of outlier-free consensus sets using fewer samplings, making the model generation more efficient. For the model selection, we propose a new Feature and Spatial consistency constrained Truncated Chamfer Distance (FS-TCD) metric for evaluating the generated models. It considers the alignment quality, the feature matching properness, and the spatial consistency constraint simultaneously, enabling the correct model to be selected even when the inlier rate of the putative correspondence set is extremely low. Extensive experiments are carried out to investigate the performance of our method. In addition, we also experimentally prove that the proposed SC $^{2}$ measure and the FS-TCD metric are general and can be easily plugged into deep learning based frameworks.

关键词： Measurement Computational modeling Estimation Point cloud compression Three-dimensional displays Deep learning Pipelines

来源：评论

学校读者我要写书评

暂无评论

An image enhancement algorithm based on multi-scale Retinex theory to improve the images quality of sensors 12

An image enhancement algorithm based on multi-scale Retinex ...

引用

SPIE 12th International Symposium on multispectral Image processing and Pattern Recognition, MIPPR 2023

作者： Li, Yiyao Chen, Zhong Sun, Wenyuan School of Journalism and Communication Central China Normal University Luoyu Road 152 Wuhan China School of Artificial Intelligence and Automation Huazhong University of Science and Technology Luoyu Road 1037 Wuhan China National Key Laboratory of Science and Technology on Multi-spectral Information Processing Luoyu Road 1037 Wuhan China Key Laboratory of Ministry of Education for Image Processing and Intelligent Control Luoyu Road 1037 Wuhan China Central China Normal University Library Luoyu Road 152 Wuhan China

ISBN: (数字)9781510674967

ISBN: (纸本)9781510674950

Nowadays, image recognition plays a pivotal role in acquiring data via sensors. However, the adaptability of traditional algorithms is hindered by the unpredictable nature of open environments, varying sensor quality, and image dimensions. Challenges arise in adverse conditions like inclement weather, low light, and optical distortions. Retinex-based methods have emerged as a viable solution, effectively enhancing images plagued by shadows or poor lighting. Yet, issues surface when images possess saturated colors;the conventional multi-scale Retinex with color restoration risks color inversion. Moreover, during gain compensation, extreme histogram values occupy significant gray level space, obscuring vital image details. This study delves into these challenges and proposes an enhanced multi-scale Retinex algorithm. Our approach substitutes logarithmic functions with tansig functions, eliminating color inversion risks. Additionally, a novel gain compensation method, integrating histogram stretching with Gamma correction, refines image clarity. The algorithm's robustness is evidenced in diverse scenarios, including adverse weather, low light, underwater imaging, and non-uniform lighting. Experimental results validate our method's superiority, surpassing other Retinex-based techniques both qualitatively and quantitatively. This research contributes valuable insights into image enhancement methodologies, fostering advancements in sensor-based data gathering in Smart Spaces. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

Beacon wavelength selection for wavefront detection in horizontal path under strong turbulence 13

Beacon wavelength selection for wavefront detection in horiz...

引用

Holography, Diffractive Optics, and Applications XIII 2023

作者： Wu, Fengyang Zhong, Sheng Wang, Xun Dong, Lizhi Liu, Ziyu Jing, Jianhang Zhang, Ying School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China National Key Laboratory of Multi-Spectral Information Intelligent Processing Huazhong University of Science and Technology Wuhan430074 China Aerospace Science & Industry Microelectronics System Institute Co. Ltd Chengdu610213 China

ISBN: (纸本)9781510667853

Atmospheric turbulence has severe influence on laser beam propagations. The well accepted method of compensating for the aberrations in the turbulence is generating a beacon on the target or at the point-ahead location. In most scenarios, the wavelength of the beacon is similar to or shorter than the launched beam. However, when the strength of the turbulence gets stronger, intensity scintillations and wavefront branch points also arise, making it difficult to accurately measure the wavefront. As scintillations and branch points weakens with wavelength increases, it is reasonable to consider using beacons with longer wavelengths to improve wavefront detection accuracy. To further investigate the feasibility of beacons with longer wavelengths, we established a simulation model of wavefront detection and correction in horizontal path under strong turbulence. A split-step multi-phase screen method is employed to simulate the propagations of laser beams. Then a series of simulations with different refractive index structure parameter Cn2 and beacon wavelengths are studied. We show that under weak turbulence, the optimal wavelength of beacon is similar to the launched beam, as is widely investigated and concluded. However, beacons with longer wavelength are preferred under strong turbulence, especially with a general wavefront reconstruction algorithm without considering branch points. This research would be helpful to for designing adaptive optics systems to improve beam qualities in strong turbulence with adaptive optics systems. © 2023 SPIE.

关键词： Atmospheric turbulence

来源：评论

学校读者我要写书评

暂无评论

A moving infrared small target detection method based on optical flow-guided neural networks

A moving infrared small target detection method based on opt...

引用

Computer Vision, Image and Deep Learning (CVIDL), International Conference on

作者： Wenming Zhu Yihua Tan National Key Laboratory of Science & Technology on Multi-spectral Information Processing School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China

In the practical application scenarios of infrared small dim target detection (IRSTD), the background is complex and changeable, which leads to poor detection performance. The traditional methods based on prior knowledge and the existing deep learning methods can only adapt to a single fixed scene, and cannot be popularized in complex and changeable scenes. Therefore, there is still a lack of a robust infrared small target detection method. In this paper, we propose an optical flow-guided neural network for moving infrared small target detection. Firstly, we design an optical flow capture module (OFCM), which can enhance the characteristics of weak targets, and capture the motion information in the image by calculating the optical flow value of the image. Secondly, we design a pyramid feature interactive network (PFI-Net) to perform feature interaction fusion of optical flow and infrared images. PFI-Net can analyse the motion information expressed by the optical flow and the contour information expressed by the infrared image on multiple scales to extract features and perform target detection and segmentation. Experimental results on public datasets show that our method can effectively detect moving targets, and has better performance than other methods.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Radio-Frequency Interference Source Detection Based on Enhanced YOLOv8 for Synthetic Aperture Interferometric Radiometers

Radio-Frequency Interference Source Detection Based on Enhan...

引用

Cross Strait Quad-Regional Radio Science and Wireless technology Conference (CSRWTC)

作者： Shimeng Lu Mingjin Yang Huanyu Wang Fanxiang Li Dong Zhu School of Electronic Information and Communications Huazhong University of Science and Technology Wuhan China National Key Laboratory of Science and Technology on Multi-Spectral Information Processing Wuhan China

The presence of radio-frequency interference (RFI) contaminates data collected by synthetic aperture interferometric radiometers (SAIRs). The performances of current RFI detection methods are limited due to insufficient exploration of RFI characteristics. This paper proposes a new RFI source detection method based on a convolutional neural network (CNN) that exploits side-lobe characteristics of the array factor of SAIR systems. Based on the common characteristic observed in RFI images, namely the presence of hexagonal tailings extended from a local peak of the RFI source, this method incorporates the hexagonal tailing feature into the detection scope, thereby expanding the capability to identify RFI. By including the hexagonal tailing feature, the method enhances the comprehensiveness and accuracy of RFI detection. This inclusion leads to a more comprehensive and accurate detection range for RFI sources. Results using real SMOS satellite data demonstrate that the proposed method has improved detection performance compared with traditional approaches that are commonly used in RFI detection tasks for SAIRs.

关键词：

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：