检索结果-内蒙古大学图书馆

Sasl: Saliency-adaptive sparsity learning for neural network acceleration

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Shi, Jun Xu, Jianfeng Tasaka, Kazuyuki Chen, Zhibo CAS Key Laboratory of Technology in Geo-spatial Information Processing and Application System University of Science and Technology of China Hefei230027 China KDDI Research Japan

Accelerating the inference speed of CNNs is critical to their deployment in real-world applications. Among all the pruning approaches, those implementing a sparsity learning framework have shown to be effective as they learn and prune the models in an end-to-end data-driven manner. However, these works impose the same sparsity regularization on all filters indiscriminately, which can hardly result in an optimal structuresparse network. In this paper, we propose a Saliency-Adaptive Sparsity Learning (SASL) approach for further optimization. A novel and effective estimation of each filter, i.e., saliency, is designed, which is measured from two aspects: The importance for the prediction performance and the consumed computational resources. During sparsity learning, the regularization strength is adjusted according to the saliency, so our optimized format can better preserve the prediction performance while zeroing out more computation-heavy filters. The calculation for saliency introduces minimum overhead to the training process, which means our SASL is very efficient. During the pruning phase, in order to optimize the proposed data-dependent criterion, a hard sample mining strategy is utilized, which shows higher effectiveness and efficiency. Extensive experiments demonstrate the superior performance of our method. Notably, on ILSVRC-201. dataset, our approach can reduce 49.7% FLOPs of ResNet-50 with very negligible 0.39% top-1.and 0.05% top-5 accuracy degradation. Copyright © 2020, The Authors. All rights reserved.

关键词： Convolutional neural networks

Static Background Imaging and Moving Target Signature Removal in Multiaspect SAR Image 6

学校读者我要写书评

暂无评论

Static Background Imaging and Moving Target Signature Remova...

6th Asia-Pacific Conference on Synthetic Aperture Radar, APSAR 201.

作者： Shen, Wenjie Hong, Wen Han, Bing Wang, Yanping Lin, Yun Key Laboratory of Technology in Geo-spatial Information Processing and Application System Beijing China Beijing China Beijing China North China University of Technology Beijing China

ISBN: (纸本)9781728129129

Multiaspect SAR has capability of providing high resolution image of static scene due to its long aperture feature. However, moving target can generate long and complex image trace in Multiaspect SAR, which may hamper the applications like SAR image interpretation and target detection. In this paper, two methods are proposed to remove moving target signature in single channel multiaspect SAR image. The two methods are all based on logarithm background subtraction. The first one is a fast scheme with cost of reduced resolution, while the second one focus on providing the high resolution but it takes more time than previous one. The first method utilize fact of target signal position changes in subaperture image sequence to obtain the static background image. The second method combines the detection results to exclude the moving target signal in each subaperture image. Then obtain the high resolution of static background by coherent summation. The methods are validated by both airborne SAR data. © 201. IEEE.

关键词： Synthetic aperture radar

Learned Image Compression with Gaussian-Laplacian-Logistic Mixture Model and Concatenated Residual Modules

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Fu, Haisheng Liang, Feng Lin, Jianping Li, Bing Akbari, Mohammad Liang, Jie Zhang, Guohe Liu, Dong Tu, Chengjie Han, Jingning The School of Microelectronics Xi’an Jiaotong University Xi’an China The CAS Key Laboratory of Technology in Geo-Spatial Information Processing and Application System University of Science and Technology of China Hefei230027 China The School of Engineering Science Simon Fraser University Canada The Tencent Technologies China The Google Inc. United States

Recently deep learning-based image compression methods have achieved significant achievements and gradually outperformed traditional approaches including the latest standard Versatile Video Coding (VVC) in both PSNR and MS-SSIM metrics. Two key components of learned image compression are the entropy model of the latent representations and the encoding/decoding network architectures. Various models have been proposed, such as autoregressive, softmax, logistic mixture, Gaussian mixture, and Laplacian. Existing schemes only use one of these models. However, due to the vast diversity of images, it is not optimal to use one model for all images, even different regions within one image. In this paper, we propose a more flexible discretized Gaussian-Laplacian-Logistic mixture model (GLLMM) for the latent representations, which can adapt to different contents in different images and different regions of one image more accurately and efficiently, given the same complexity. Besides, in the encoding/decoding network design part, we propose a concatenated residual blocks (CRB), where multiple residual blocks are serially connected with additional shortcut connections. The CRB can improve the learning ability of the network, which can further improve the compression performance. Experimental results using the Kodak, Tecnick-1.0 and Tecnick-40 datasets show that the proposed scheme outperforms all the leading learning-based methods and existing compression standards including VVC intra coding (4:4:4 and 4:2:0) in terms of the PSNR and MS-SSIM. The source code is available at https://***/fengyurenpingsheng. Copyright © 2021. The Authors. All rights reserved.

关键词： Image compression

Cascaded Regression Tracking: Towards Online Hard Distractor Discrimination

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Wang, Ning Zhou, Wengang 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 Hefei China Huawei Noah’s Ark Laboratory

Visual tracking can be easily disturbed by similar surrounding objects. Such objects as hard distractors, even though being the minority among negative samples, increase the risk of target drift and model corruption, which deserve additional attention in online tracking and model update. To enhance the tracking robustness, in this paper, we propose a cascaded regression tracker with two sequential stages. In the first stage, we filter out abundant easily-identified negative candidates via an efficient convolutional regression. In the second stage, a discrete sampling based ridge regression is designed to double-check the remaining ambiguous hard samples, which serves as an alternative of fully-connected layers and benefits from the closed-form solver for efficient learning. During the model update, we utilize the hard negative mining technique and an adaptive ridge regression scheme to improve the discrimination capability of the second-stage regressor. Extensive experiments are conducted on 1. challenging tracking benchmarks including OTB-201., OTB-201., VOT201., VOT201., UAV1.3, Temple-Color, NfS, TrackingNet, LaSOT, UAV20L, and OxUvA. The proposed method achieves state-of-the-art performance on prevalent benchmarks, while running in a real-time speed. Copyright © 2020, The Authors. All rights reserved.

关键词： Regression analysis

Stereo Matching of VHR Remote Sensing Images via Bidirectional Pyramid Network

学校读者我要写书评

暂无评论

Stereo Matching of VHR Remote Sensing Images via Bidirection...

IEEE International Symposium on geoscience and Remote Sensing (IGARSS)

作者： Rongshu Tao Yuming Xiang Hongjian You Key Laboratory of Technology in Geo-spatial Information Processing and Application System Aerospace Information Research Institute Chinese Academy of Science Beijing China University of Chinese Academy of Sciences Beijing China

ISBN: (数字)9781728163741

ISBN: (纸本)9781728163758

Stereo matching, an essential step in 3D reconstruction, still faces unignorable problems due to the very high resolution and complex structures of remote sensing images. Especially in occluded areas of high buildings and untextured areas of waters and woods, precise disparity estimation has become a difficult but important task. In this paper, we propose a novel method based on the pyramid stereo matching network to solve the aforementioned problems. Inspired by the classical optical flow estimation framework, we adopt the forward-backward consistency assumption to improve the accuracy. Moreover, we improve the construction of cost volume since the traditional deep-learning networks only work well for positive disparities and the disparity ranges in remote sensing images vary a lot. The proposed network is compared with two baselines. The experimental results show that our proposed method outperforms two baselines in terms of average endpoint error (EPE) and the fraction of erroneous pixels(D1., and the improvements in occluded areas are significant.

关键词： Estimation Remote sensing Optical sensors Optical imaging Three-dimensional displays Buildings Satellites

Sequential reinforced 360-degree video adaptive streaming with cross-user attentive network

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Fu, Jun Chen, Zhibo Chen, Xiaoming Li, Weiping CAS Key Laboratory of Technology in Geo-spatial Information Processing and Application System University of Science and Technology of China Hefei230027 China School of Computer Science and Engineering Beijing Technology and Business University China

In the tile-based 360-degree video streaming, predicting user’s future viewpoints and developing adaptive bitrate (ABR) algorithms are essential for optimizing user’s quality of experience (QoE). Traditional single-user based viewpoint prediction methods fail to achieve good performance in long-term prediction, and the recently proposed reinforcement learning (RL) based ABR schemes applied in traditional video streaming can not be directly applied in the tile-based 360-degree video streaming due to the exponential action space. Therefore, we propose a sequential reinforced 360-degree video streaming scheme with cross-user attentive network. Firstly, considering different users may have the similar viewing preference on the same video, we propose a cross-user attentive network (CUAN), boosting the performance of long-term viewpoint prediction by selectively utilizing cross-user information. Secondly, we propose a sequential RL-based (360SRL) ABR approach, transforming action space size of each decision step from exponential to linear via introducing a sequential decision structure. We evaluate the proposed CUAN and 360SRL using trace-driven experiments and experimental results demonstrate that CUAN and 360SRL outperform existing viewpoint prediction and ABR approaches with a noticeable margin. Copyright © 2020, The Authors. All rights reserved.

关键词： Forecasting

Phase Transition Analysis for Compressive Sensing Based DoA Estimation in Automotive Radar

学校读者我要写书评

暂无评论

Phase Transition Analysis for Compressive Sensing Based DoA ...

International Radar Symposium (IRS)

作者： Yan Zhang Silin Gao Mingqian Liu Bingchen Zhang Yirong Wu Chinese Academy of Sciences Aerospace Information Research Institute Beijing China University of Chinese Academy of Sciences Beijing China Key Laboratory of Technology in Geo-spatial Information and Processing Application System Beijing China

ISBN: (数字)9788394942151

ISBN: (纸本)9781728157863

In automotive radar applications, the compressive sensing (CS) based DoA estimation is used in array signal {1. in recent years. Sparse reconstruction has the potential to estimate the direction of arrival (DoA) with super resolution. However, failed results may be acquired via sparse reconstruction in inappropriate conditions, namely the critical condition determining success or failure must be taken into consideration. In this paper, the sparsity of the scenario and the signal-to-noise ratio (SNR) are analyzed as the main factors via phase transition diagrams. Other factors affecting the success or failure are also investigated, such as the array configuration and the sparse recovery algorithm. Simulated and experimental results demonstrate the critical conditions, in which the DoA estimation is successful or failed.

关键词： Direction-of-arrival estimation Estimation Radar applications Compressed sensing Signal resolution Signal to noise ratio Automotive engineering

On the Use of Axial Ratio to Analysis Transmit Distortions in Hybrid Polarimetry and Their Influence to Canonical Scattering Mechanisms 6

学校读者我要写书评

暂无评论

On the Use of Axial Ratio to Analysis Transmit Distortions i...

6th Asia-Pacific Conference on Synthetic Aperture Radar, APSAR 201.

作者： Chen, Shiqiang Hong, Wen Key Laboratory of Technology in Geo-spatial Information Processing and Application System Beijing China Institute of Electronics Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China

ISBN: (纸本)9781728129129

Transmit distortions in the hybrid polarimetric (HP) SAR cannot be compensated simply with external calibration methods. Therefore, it is necessary to analysis their influence on HP data. In this study, we have proposed an analysis method based on the Axial Ratio (AR) of the real transmitted polarized wave. Firstly, impact of different distortion sources is quantified with the AR parameter. Next, a case study is illustrated using real GF-3 measured data. At last, the influence of transmit distortions to canonical scattering mechanisms is analyzed. © 201. IEEE.

关键词： Polarimeters

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 {1., 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