检索结果-内蒙古大学图书馆

DeepEraser: Deep Iterative Context Mining for Generic Text Eraser

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Feng, Hao Wang, Wendi Liu, Shaokai Deng, Jiajun Zhou, Wengang Li, Houqiang The CAS Key Laboratory of Technology in Geo-spatial Information Processing and Application System Department of Electronic Engineering and Information Science University of Science and Technology of China Hefei230027 China The University of Adelaide Australian Institute for Machine Learning Australia

In this work, we present DeepEraser, an effective deep network for generic text removal. DeepEraser utilizes a recurrent architecture that erases the text in an image via iterative operations. Our idea comes from the process of erasing pencil script, where the text area designated for removal is subject to continuous monitoring and the text is attenuated progressively, ensuring a thorough and clean erasure. Technically, at each iteration, an innovative erasing module is deployed, which not only explicitly aggregates the previous erasing progress but also mines additional semantic context to erase the target text. Through iterative refinements, the text regions are progressively replaced with more appropriate content and finally converge to a relatively accurate status. Furthermore, a custom mask generation strategy is introduced to improve the capability of DeepEraser for adaptive text removal, as opposed to indiscriminately removing all the text in an image. Our DeepEraser is notably compact with only 1.4M parameters and trained in an end-to-end manner. To verify its effectiveness, extensive experiments are conducted on several prevalent benchmarks, including SCUT-Syn, SCUT-EnsText, and Oxford Synthetic text dataset. The quantitative and qualitative results demonstrate the effectiveness of our DeepEraser over the state-of-the-art methods, as well as its strong generalization ability in custom mask text removal. The codes and pre-trained models are available at https://***/fh2019ustc/DeepEraser Copyright © 2024, The Authors. All rights reserved.

关键词： Iterative methods

A space-based radar nonlinear tracking method for air targets based on VB-PHD filter

学校读者我要写书评

暂无评论

A space-based radar nonlinear tracking method for air target...

Signal, information and Data processing (ICSIDP), IEEE International Conference on

作者： Linlin Fang Lihua Zhong Bing Han Yuxin Hu Aerospace Information Research Institute Chinese Academy of Sciences Key Laboratory of Technology in Geo-Spatial Information Processing and Application Systems Chinese Academy of Sciences Beijing China

ISBN: (数字)9798331515669

ISBN: (纸本)9798331515676

Space-based radar has better coverage and is not limited by the curvature of the earth. Aircraft target tracking is particularly important in space-based radar system. However, the long observation distance and strong nonlinear observation increases the challenge of target tracking. Therefore, we proposed a space-based radar nonlinear tracking method for air targets based on variational Bayesian probability hypothesis density (VB-PHD) filter. The relationship between state and observation is established. Then, the VB method is adopted to deal with uncertainty of measurement noise variances. Besides, the cubature filter is introduced to deal with the strongly nonlinear models, and the recursion process is completed by using the third-order spherical radial volume rule. Finally, simulation scenarios with different measurement noise parameters were constructed, and the tracking result comparisons show that it effectively solves nonlinear problems, and tracking precision can be significantly improved.

关键词： Solid modeling Target tracking Uncertainty Spaceborne radar Simulation Noise Measurement uncertainty Radar tracking Bayes methods Noise measurement

Study on Pixel Level Segmentation and Area Quantification of Highway Slope Cracks 3rd

学校读者我要写书评

暂无评论

Study on Pixel Level Segmentation and Area Quantification of...

3rd EAI International Conference on Artificial Intelligence for Communications and Networks, AICON 2021

作者： Zhang, YunLing Guo, Pei Liu, Pengyu Li, Yaoyao Chen, Shanji Research and Development Center of Transport Industry of Spatial Information Application and Disaster Prevention and Mitigation Technology Beijing100097 China China Highway Engineering Consultants Corporation Beijing100097 China Faculty of Information Technology Beijing University of Technology Beijing100124 China Beijing Laboratory of Advanced Information Networks Beijing100124 China Beijing Key Laboratory of Computational Intelligence and Intelligent System Beijing University of Technology Beijing100124 China School of Physics and Electronic Information Qinghai Nationalities University Xining810007 China

ISBN: (纸本)9783030901950

Highway slope disasters show obvious stage characteristics before the occurrence, and cracks are the early symptoms of most highway slope disasters. Computer vision is widely used in crack detection because of its advantages of high efficiency and low cost. In view of the shortage of traditional crack actual area calculation methods and poor effect, this paper proposes a slope crack pixel level segmentation method based on deep convolutional neural network, so as to generate accurate segmentation of crack morphology. Then, according to the binary segmentation mask, the checkerboard mapping method is proposed to calculate the actual crack area. Finally, the effectiveness and superiority of the proposed checkerboard mapping method are verified and evaluated with a self-made data set of highway crack image. The experimental results show that this method can effectively detect the actual crack area, and the relative error is small. The calculated results can be used as a reference for slope disaster warning. © 2021, ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering.

关键词： Deep neural networks

ROAD EXTRACTION FROM REMOTE SENSING IMAGES BY MULTIPLE FEATURE PYRAMID NETWORK

学校读者我要写书评

暂无评论

ROAD EXTRACTION FROM REMOTE SENSING IMAGES BY MULTIPLE FEATU...

IEEE International Geoscience and Remote Sensing Symposium

作者： Xun Gao Xian Sun Menglong Yan Hao Sun Kun Fu Yue Zhang Zhipeng Ge School of Electronic Electrical and Communication Engineering University of Chinese Academy of Sciences Beijing China Key Laboratory of Technology in Geo-spatial Information Processing and Application System Chinese Academy of Sciences Beijing China

Road extraction from high-resolution remote sensing images has been applied in many domains, but it is still full of challenges. We focus on the problem of slender roads, proposing a new multiple feature pyramid network (MFPN), which is composed of an effective feature pyramid and the tailored pyramid pooling module based on PSPNet. These two designs can address the sparsity of roads in remote sensing images via using multi-level semantic features. Experiments on remote sensing images from Quick Bird show that our MFPN model achieves competitive performance, especially for slender roads.

关键词： Roads Feature extraction Remote sensing Convolution Semantics Image segmentation Data mining

A HYBRID NEURAL NETWORK FOR CHROMA INTRA PREDICTION

学校读者我要写书评

暂无评论

A HYBRID NEURAL NETWORK FOR CHROMA INTRA PREDICTION

IEEE International Conference on Image processing

作者： Yue Li Li Li Zhu Li Jianchao Yang Ning Xu Dong Liu Houqiang Li CAS Key Laboratory of Technology in Geo-Spatial Information Processing and Application System University of Science and Technology of China Hefei China University of Missouri Kansas City MO USA Snapchat Inc. CA USA

For chroma intra prediction, previous methods exemplified by the Linear Model method (LM) usually assume a linear correlation between the luma and chroma components in a coding block. This assumption is inaccurate for complex image content or large blocks, and restricts the prediction accuracy. In this paper, we propose a chroma intra prediction method by exploiting both spatial and cross-channel correlations using a hybrid neural network. Specifically, we utilize a convolutional neural network to extract features from the reconstructed luma samples of the current block, as well as utilize a fully connected network to extract features from the neighboring reconstructed luma and chroma samples. The extracted twofold features are then fused to predict the chroma samples-Cb and Cr simultaneously. The proposed chroma intra prediction method is integrated into HEVC. Preliminary results show that, compared with HEVC plus LM, the proposed method achieves on average 0.2%, 3.1% and 2.0% BD-rate reduction on Y, Cb and Cr components, respectively, under All-Intra configuration.

关键词： Feature extraction Neural networks Kernel Correlation Image reconstruction Encoding Prediction methods

Progressive unsupervised person re-identification by tracklet association with spatio-temporal regularization

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Xie, Qiaokang Zhou, Wengang Qi, Guo-Jun Tian, Qi Li, Houqiang CAS Key Laboratory of Technology in Geo-spatial Information Processing and Application System Department of Electronic Engineering and Information Science University of Science and Technology of China Hefei230027 China Huawei Cloud EI Product Department Huawei Noah's Ark Laboratory

Existing methods for person re-identification (Re-ID) are mostly based on supervised learning which requires numerous manually labeled samples across all camera views for training. Such a paradigm suffers the scalability issue since in real-world Re-ID application, it is difficult to exhaustively label abundant identities over multiple disjoint camera views. To this end, we propose a progressive deep learning method for unsupervised person Re-ID in the wild by Tracklet Association with Spatio-Temporal Regularization (TASTR). In our approach, we first collect tracklet data within each camera by automatic person detection and tracking. Then, an initial Re-ID model is trained based on within-camera triplet construction for person representation learning. After that, based on the person visual feature and spatio-temporal constraint, we associate cross-camera tracklets to generate cross-camera triplets and update the Re-ID model. Lastly, with the refined Re-ID model, better visual feature of person can be extracted, which further promote the association of cross-camera tracklets. The last two steps are iterated multiple times to progressively upgrade the Re-ID model. To facilitate the study, we have collected a new 4K UHD video dataset named Campus-4K with full frames and full spatio-temporal information. Experimental results show that with the spatio-temporal constraint in the training phase, the proposed approach outperforms the state-of-the-art unsupervised methods by notable margins on DukeMTMC-reID, and achieves competitive performance to fully supervised methods on both DukeMTMC-reID and Campus-4K datasets. Copyright © 2019, The Authors. All rights reserved.

关键词： Cameras

A3Net: Adversarial-and-attention network for machine reading comprehension

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Wang, Jiuniu Fu, Xingyu Xu, Guangluan Wu, Yirong Chen, Ziyan Wei, Yang Jin, Li Key Laboratory of Technology in Geo-spatial Information Processing and Application System Institute of Electronics CAS Beijing China School of Electronic Electrical and Communication Engineering University of Chinese Academy of Sciences Beijing China

In this paper, we introduce Adversarial-and-attention Network (A3Net) for Machine Reading Comprehension. This model extends existing approaches from two perspectives. First, adversarial training is applied to several target variables within the model, rather than only to the inputs or embeddings. We control the norm of adversarial perturbations according to the norm of original target variables, so that we can jointly add perturbations to several target variables during training. As an effective regularization method, adversarial training improves robustness and generalization of our model. Second, we propose a multi-layer attention network utilizing three kinds of high-efficiency attention mechanisms. Multi-layer attention conducts interaction between question and passage within each layer, which contributes to reasonable representation and understanding of the model. Combining these two contributions, we enhance the diversity of dataset and the information extracting ability of the model at the same time. Meanwhile, we construct A3Net for the WebQA dataset. Results show that our model outperforms the state-of-the-art models (improving Fuzzy Score from 73.50% to 77.0%). Copyright © 2018, The Authors. All rights reserved.

关键词： Network layers

Prior-enlightened and motion-robust video deblurring

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Zhou, Ya Xu, Jianfeng Tasaka, Kazuyuki Chena, Zhibo Li, Weiping CAS Key Laboratory of Technology in Geo-Spatial Information Processing and Application System University of Science and Technology of China Hefei230026 China KDDI Research Inc. Fujimino-shi Saitama356-8502 Japan

Various blur distortions in video will cause negative impact on both human viewing and video-based applications, which makes motion-robust deblurring methods urgently needed. Most existing works have strong dataset dependency and limited generalization ability in handling challenging scenarios, like blur in low contrast or severe motion areas, and non-uniform blur. Therefore, we propose a PRiOr-enlightened and MOTION-robust video deblurring model (PROMOTION) suitable for challenging blurs. On the one hand, we use 3D group convolution to efficiently encode heterogeneous prior information, explicitly enhancing the scenes’ perception while mitigating the output’s artifacts. On the other hand, we design the priors representing blur distribution, to better handle non-uniform blur in spatio-temporal domain. Besides the classical camera shake caused global blurry, we also prove the generalization for the downstream task suffering from local blur. Extensive experiments demonstrate we can achieve the state-of-the-art performance on well-known REDS and GoPro datasets, and bring machine task gain. Copyright © 2020, The Authors. All rights reserved.

关键词： Motion estimation

DENSE-ADD NET: AN NOVEL CONVOLUTIONAL NEURAL NETWORK FOR REMOTE SENSING IMAGE INPAINTING

学校读者我要写书评

暂无评论

DENSE-ADD NET: AN NOVEL CONVOLUTIONAL NEURAL NETWORK FOR REM...

IEEE International Geoscience and Remote Sensing Symposium

作者： Daoyu Lin Guangluan Xu Yang Wang Xian Sun Kun Fu School of Electronic Electrical and Communication Engineering University of Chinese Academy of Sciences Beijing China Key Laboratory of Technology in Geo-spatial Information Processing and Application System Chinese Academy of Sciences Beijing China

Through the recent performance of convolutional neural networks in image processing tasks, we propose a deep fully convolutional network for remote sensing image inpainting. The proposed Dense-Add Net (Dense-Add Network) can alleviate the vanishing-gradient problem, strengthen feature reuse, and substantially reduce the memory usage. We apply residual learning to learn the mappings from corrupted image to recovered image directly;it will back-propagate gradient to the bottom layers and accelerate the training process. We train the proposed Dense-Add Net with a robust Charbonnier loss function which can achieve high-quality reconstruction. The experimental verify the efficacy of our proposed Dense-Add Net.

关键词： Remote sensing Training Acceleration Image reconstruction Painting Convolution Interpolation