检索结果-内蒙古大学图书馆

ieee computer society conference on computer vision and pattern recognition workshops (CVPRW)

作者： Yingqian Wang Zhengyu Liang Qianyu Chen Longguang Wang Jungang Yang Radu Timofte Yulan Guo Wentao Chao Yiming Kan Xuechun Wang Fuqing Duan Guanghui Wang Wang Xia Ziqi Wang Yue Yan Peiqi Xia Shunzhou Wang Yao Lu Angulia Yang Kai Jin Zeqiang Wei Sha Guo Mingzhi Gao Xiuzhuang Zhou Zhongxin Yu Shaofei Luo Cheng Zhong Shaorui Chen Long Peng Yuhong He Gaosheng Liu Huanjing Yue Jingyu Yang Zhengjian Yao Jiakui Hu Lujia Jin Zhi-Song Liu Chenhang He Jun Xiao Xiuyuan Wang Zonglin Tian Yifan Mao Deyang Liu Shizheng Li Ping An

ISBN: (数字)9798350365474

ISBN: (纸本)9798350365481

In this report, we summarize the 2nd NTIRE challenge on light field (LF) image super-resolution (SR) with a focus on new methods and results. This challenge aims at superresolving LF images under the standard bicubic downsampling degradation with a magnification factor of ×4. Compared with single image SR, the major challenge of LF image SR lies in how to exploit complementary angular information from plenty of views with varying disparities. This year of challenge has two tracks, including one track on fidelity (i.e., restoration accuracy in terms of PSNR) only, and the other track on fidelity with an extra constraint on model size and computational cost. In total, 125 participants were successfully registered for this challenge, and 9 teams have successfully submitted results with PSNR scores higher than the baseline methods. We report the solutions proposed by the participants, and summarize their common trends and useful tricks. We hope this challenge can stimulate future research and inspire new ideas in LF image SR.

关键词： Degradation computer vision conferences Computational modeling Superresolution Market research Light fields

来源：评论

学校读者我要写书评

暂无评论

Unsupervised Domain Adaptation Architecture Search with Self-Training for Land Cover Mapping

Unsupervised Domain Adaptation Architecture Search with Self...

引用

ieee computer society conference on computer vision and pattern recognition workshops (CVPRW)

作者： Clifford Broni-Bediako Junshi Xia Naoto Yokoya RIKEN RIKEN Center for Advanced Intelligence Project (AIP) Tokyo Japan Graduate School of Frontier Sciences The University of Tokyo Chiba Japan

ISBN: (数字)9798350365474

ISBN: (纸本)9798350365481

Unsupervised domain adaptation (UDA) is a challenging open problem in land cover mapping. Previous studies show encouraging progress in addressing cross-domain distribution shifts on remote sensing benchmarks for land cover mapping. The existing works are mainly built on large neural network architectures, which makes them resource-hungry systems, limiting their practical impact for many real-world applications in resource-constrained environments. Thus, we proposed a simple yet effective framework to search for lightweight neural networks automatically for land cover mapping tasks under domain shifts. This is achieved by integrating Markov random field neural architecture search (MRF-NAS) into a self-training UDA framework to search for efficient and effective networks under a limited computation budget. This is the first attempt to combine NAS with self-training UDA as a single framework for land cover mapping. We also investigate two different pseudo-labelling approaches (confidence-based and energy-based) in self-training scheme. Experimental results on two recent datasets (OpenEarthMap & FLAIR #1) for remote sensing UDA demonstrate a satisfactory performance. With only less than 2M parameters and 30.16 G FLOPs, the best-discovered lightweight network reaches state-of-the-art performance on the regional target domain of OpenEarthMap (59.38% mIoU) and the considered target domain of FLAIR #1 (51.19% mIoU). The code is at https://***/cliffbb/UDA-NAS.

关键词： Limiting conferences Neural networks Land surface computer architecture Benchmark testing Search problems

来源：评论

学校读者我要写书评

暂无评论

Wake-Sleep Energy Based Models for Continual Learning

Wake-Sleep Energy Based Models for Continual Learning

引用

ieee computer society conference on computer vision and pattern recognition workshops (CVPRW)

作者： Vaibhav Singh Anna Choromanska Shuang Li Yilun Du Concordia University Mila New York University University of Toronto Massachusetts Institute of Technology

ISBN: (数字)9798350365474

ISBN: (纸本)9798350365481

This paper introduces a novel approach for continually training Energy-Based Models (EBMs) on the classification problems in the challenging setting of class incremental learning. Despite the fact that EBMs offer longer retention of knowledge on prior tasks, training EBMs contrastively remains a challenge. Driven by biological plausibility, we leverage the observation that sleep in humans supports active system consolidation and propose a new approach for training EBMs, which we call Wake-Sleep Energy Based Models (WS-EBMs), which rely on wake-sleep cycles. Our training approach consists of short wake phases followed by long sleep phases. During the short wake phase, the free energy associated with ground truth labels is minimized, which conditions the model towards the correct solutions. This is followed by a long sleep phase, where the free energy of the whole system is minimized contrastively, which allows the model to push the energy of incorrect solutions further from the correct response. We provide a theoretical analysis of WS-EBM showing that it satisfies the sufficient condition for designing proper EBM loss. Our empirical evaluation confirms the plausibility of our approach and demonstrates favorable performance of WS-EBM compared to traditional EBM training as well as state-of-the-art class-incremental continual learning techniques. Furthermore, our proposed two-phase training strategy can be easily integrated with existing techniques resulting in substantial boosts in their performance. Finally, we also provide interesting insights justifying our approach by analyzing the orthogonality between the sequential task vectors, and flatness of the optimized energy surfaces, which may guide the design of class incremental continual learning strategies.

关键词： Training Continuing education Sufficient conditions computer vision Incremental learning Biological system modeling conferences

来源：评论

学校读者我要写书评

暂无评论

RLNet: Robust Linearized Networks for Efficient Private Inference

RLNet: Robust Linearized Networks for Efficient Private Infe...

引用

ieee computer society conference on computer vision and pattern recognition workshops (CVPRW)

作者： Sreetama Sarkar Souvik Kundu Peter A. Beereil Universiy of Southern California Los Angeles USA Intel Labs San Diego USA

ISBN: (数字)9798350365474

ISBN: (纸本)9798350365481

The growing concern about data privacy has led to the development of private inference (PI) frameworks in client-server applications which protects both data privacy and model IP. However, the cryptographic primitives required yield significant latency overheads which limits their wide-spread application. At the same time, changing environments demand PI services to be robust against various naturally occurring and gradient-based perturbations. Despite several works focused on the development of latency-efficient models suitable for PI, the impact of these models on robustness has remained unexplored. Towards this goal, this paper presents RLNet, a class of models that can yield latency improvement via the reduction of high-latency ReLU operations while improving the model performance on both clean and corrupted images. In particular, RLNet models provide a "triple win ticket" of improved classification accuracy on clean, naturally perturbed, and gradient- based perturbed images using a shared-mask shared-weight architecture with over an order of magnitude fewer ReLUs than baseline models. To demonstrate the efficacy of RLNet, we perform extensive experiments with ResNet and WRN model variants on CIFAR-10, CIFAR-100, and Tiny- ImageNet datasets. Our experimental evaluations show that RLNet can yield models with up to 11.14× fewer ReLUs, with accuracy close to the all-ReLU models, on clean, naturally perturbed, and gradient-based perturbed images. Compared with the SoTA non-robust linearized models at similar ReLU budgets, RLNet achieves an improvement in adversarial accuracy of up to ~47%, in naturally perturbed accuracy of up to ~16.4%, while improving clean image accuracy up to ~1.5%. Code is available at: https://***/sreetamasarkar/rlnet.

关键词： Data privacy computer vision Accuracy Computational modeling Perturbation methods Transformers Robustness

来源：评论

学校读者我要写书评

暂无评论

Color-cued Efficient Densification Method for 3D Gaussian Splatting

Color-cued Efficient Densification Method for 3D Gaussian Sp...

引用

ieee computer society conference on computer vision and pattern recognition workshops (CVPRW)

作者： Sieun Kim Kyungjin Lee Youngki Lee Seoul National University Republic of Korea

ISBN: (数字)9798350365474

ISBN: (纸本)9798350365481

Many variants of Neural Radiance Fields (NeRF) have been explored in pursuit of high-quality results with reasonable data size and real-time rendering speed. 3D Gaussian Splatting (3DGS) gained popularity due to its ability to render quality images in real-time; however, it still faces challenges with large data sizes. Meanwhile, the densification process of 3DGS plays a large role in deciding the quality and the data size of a model. Hence, it is crucial to devise a densification method that can populate Gaussians efficiently so that quality can be enhanced and fewer Gaussians are used. An efficient densification method that results in fewer Gaussians can also promote efficiency in training time, GPU memory usage, and rendering ***, we propose a novel, efficient densification method based on color cues, aiming to achieve a more compact Gaussian model without sacrificing image quality. By expanding the original 3DGS densification scheme, we identify weaknesses in the original method that lead to redundant Gaussians and compromise quality. In contrast to the original approach, which relies solely on the 2D position gradient, our method additionally leverages the spherical harmonics (SH) gradient to consider color cues. This approach resolves the inefficiencies of the original densification by aligning with the expanded scheme. Our method achieves at least 9× data size reduction with increased perceptual quality, accompanied by additional efficiencies in training time, GPU memory usage, and rendering speed.

关键词： Training Image quality Three-dimensional displays Image color analysis Memory management Graphics processing units Rendering (computer graphics)

来源：评论

学校读者我要写书评

暂无评论

Correction to: Machine translation from signed to spoken languages: state of the art and challenges (Universal Access in the Information society, (2023), 10.1007/s10209-023-00992-1)

引用

Universal Access in the Information society 2024年第1期24卷 953-957页

作者： De Coster, Mathieu Shterionov, Dimitar Van Herreweghe, Mieke Dambre, Joni IDLab-AIRO Ghent University - imec Technologiepark-Zwijnaarde 126 Ghent9052 Belgium Tilburg University Warandelaan 2 Tilburg5037 AB Netherlands Ghent University Blandijnberg 2 Ghent9000 Belgium

In this article the incorrect Fig. 7 appear as below: (Figure presented.) The Fig. 7 should have appeared as shown below. (Figure presented.) Transformers tend to outperform RNNs on different SLT tasks in terms of BLEU-4 score on the RWTH-PHOENIX-Weather 2014T dataset In this article the incorrect Fig. 8 appear as below: (Figure presented.) The Fig. 8 should have appeared as shown below. (Figure presented.) Evolution of model scores on the RWTH-PHOENIX-Weather 2014T per task In this article the incorrect Table 6 appear as below: (Figure presented.) The Table 6 should have appeared as shown below. (Figure presented.) The Table 7 appears incorrectly as below. (Figure presented.) The correct Table 7 should be shown as below. (Figure presented.) This erratum contains changes to the findings from two of the cited papers. For citations [88] and [99]: [88] Chen, Y., Zuo, R., Wei, F., Wu, Y., Liu, S., & Mak, B.: Two-stream network for sign language recognition and translation. Advances in Neural Information Processing Systems, 35, 17043–17056 (2022) [99] Chen, Y., Wei, F., Sun, X., Wu, Z., Lin, S.: A simple multi-modality transfer learning baseline for sign language translation. In: Proceedings of the ieee/CVF conference on computer vision and pattern recognition, pp. 5120–5130 (2022) The original article has been corrected. © 2024, Springer-Verlag GmbH Germany, part of Springer Nature.

关键词：

来源：评论

学校读者我要写书评

暂无评论

T2LM: Long-Term 3D Human Motion Generation from Multiple Sentences

T2LM: Long-Term 3D Human Motion Generation from Multiple Sen...

引用

ieee computer society conference on computer vision and pattern recognition workshops (CVPRW)

作者： Taeryung Lee Fabien Baradel Thomas Lucas Kyoung Mu Lee Grégory Rogez IPAI & ASRI Naver LABS Europe Dept. of ECE Seoul National University Korea

ISBN: (数字)9798350365474

ISBN: (纸本)9798350365481

In this paper, we address the challenging problem of long-term 3D human motion generation. Specifically, we aim to generate a long sequence of smoothly connected actions from a stream of multiple sentences (i.e., paragraph). Previous long-term motion generating approaches were mostly based on recurrent methods, using previously generated motion chunks as input for the next step. However, this approach has two drawbacks: 1) it relies on sequential datasets, which are expensive; 2) these methods yield unrealistic gaps between motions generated at each step. To address these issues, we introduce simple yet effective T2LM, a continuous long-term generation framework that can be trained without sequential data. T2LM comprises two components: a 1D-convolutional VQVAE, trained to compress motion to sequences of latent vectors, and a Transformer-based Text Encoder that predicts a latent sequence given an input text. At inference, a sequence of sentences is translated into a continuous stream of latent vectors. This is then decoded into a motion by the VQVAE decoder; the use of 1D convolutions with a local temporal receptive field avoids temporal inconsistencies between training and generated sequences. This simple constraint on the VQ-VAE allows it to be trained with short sequences only and produces smoother transitions. T2LM outperforms prior long-term generation models while overcoming the constraint of requiring sequential data; it is also competitive with SOTA single-action generation models.

关键词： Training computer vision Three-dimensional displays conferences Computational modeling Transformers Vectors

来源：评论

学校读者我要写书评

暂无评论

Test-time Assessment of a Model’s Performance on Unseen Domains via Optimal Transport

Test-time Assessment of a Model’s Performance on Unseen Dom...

引用

ieee computer society conference on computer vision and pattern recognition workshops (CVPRW)

作者： Akshay Mehra Yunbei Zhang Jihun Hamm Tulane University

ISBN: (数字)9798350365474

ISBN: (纸本)9798350365481

Gauging the performance of ML models on data from unseen domains at test-time is essential yet a challenging problem due to the lack of labels in this setting. Moreover, the performance of these models on in-distribution data is a poor indicator of their performance on data from unseen domains. Thus, it is essential to develop metrics that can provide insights into the model’s performance at test time and can be computed only with the information available at test time (such as their model parameters, the training data or its statistics, and the unlabeled test data). To this end, we propose a metric based on Optimal Transport that is highly correlated with the model’s performance on unseen domains and is efficiently computable only using information available at test time. Concretely, our metric characterizes the model’s performance on unseen domains using only a small amount of unlabeled data from these domains and data or statistics from the training (source) domain(s). Through extensive empirical evaluation using standard benchmark datasets, and their corruptions, we demonstrate the utility of our metric in estimating the model’s performance in various practical applications. These include the problems of selecting the source data and architecture that leads to the best performance on data from an unseen domain and the problem of predicting a deployed model’s performance at test time on unseen domains. Our empirical results show that our metric, which uses information from both the source and the unseen domain, is highly correlated with the model’s performance, achieving a significantly better correlation than that obtained via the popular prediction entropy-based metric, which is computed solely using the data from the unseen domain.

关键词： Measurement Training Correlation Computational modeling Training data computer architecture Predictive models

来源：评论

学校读者我要写书评

暂无评论

Motorcyclist Helmet Violation Detection Framework by Leveraging Robust Ensemble and Augmentation Methods

Motorcyclist Helmet Violation Detection Framework by Leverag...

引用

ieee computer society conference on computer vision and pattern recognition workshops (CVPRW)

作者： Thien Van Luong Huu Si Phuc Nguyen Duy Khanh Dinh Viet Hung Duong Duy Hong Sam Vo Huan Vu Minh Tuan Hoang Tien Cuong Nguyen Faculty of Computer Science Phenikaa University Hanoi Vietnam VNPT Group VNPT AI Hanoi Vietnam University of Transport and Communications Hanoi Vietnam

ISBN: (数字)9798350365474

ISBN: (纸本)9798350365481

Traffic Monitoring Systems play a crucial role in real-life scenarios by improving traffic flow and reducing violations. Among these violations, helmet non-compliance is particularly common in countries where motorcycles are the primary mode of transportation. However, deploying an automatic violation capturing for helmet non-compliance in the real world presents challenges due to diverse traffic situations, environmental conditions, varying object sizes, and severely imbalanced datasets. In order to address these challenges, we propose a novel deep learning framework for helmet violation detection, which consists of four different modules, namely, object detection, object association, post-processing for tracking, and score correction. In particular, we develop a robust ensemble method to take advantage of various state-of-the-art object detection models such as YOLOv7, YOLOv8, Co-DETR, and EfficientDet. Furthermore, to address the issue of data imbalance, we propose two copy and paste data augmentation techniques for enriching data samples of rare classes. As a result, our approach yields a substantial 7.43% mAP enhancement over the baseline Co-DETR model, achieving a final score of 0.4792 in the 2024 AI City Challenge Track 5 test set and ranking 3rd among the competing teams.

关键词： Deep learning Head Urban areas Transportation Motorcycles Object detection Data augmentation

来源：评论

学校读者我要写书评

暂无评论

CoLa-SDF: Controllable Latent StyleSDF for Disentangled 3D Face Generation

CoLa-SDF: Controllable Latent StyleSDF for Disentangled 3D F...

引用

ieee computer society conference on computer vision and pattern recognition workshops (CVPRW)

作者： Rahul Dey Bernhard Egger Vishnu Naresh Boddeti Ye Wang Tim K. Marks Mitsubishi Electric Research Laboratories (MERL) Michigan State University (MSU) Friedrich-Alexander-Universitat (FAU) Erlangen-Nürnberg

ISBN: (数字)9798350365474

ISBN: (纸本)9798350365481

Generating 3D faces and rendering them to images has numerous practical applications in areas including AR/VR, dataset generation, and avatar creation. In recent years, there has been a significant surge in the development of high-fidelity 3D face generation techniques such as StyleSDF, which combine the benefits of 3D implicit neural representations with those of style-based 2D generative adversarial networks (GANs). Although these implicit 3D GAN approaches generate highly realistic faces using a 3D representation, the properties of the generated faces cannot easily be edited or controlled. Meanwhile, linear 3D morphable models (3DMMs) and their nonlinear extensions have also made significant strides in their expressive capacity and quality, but they have yet to match the image quality achieved by GANs. This paper proposes a new method, CoLa-SDF, which combines the controllability of nonlinear 3DMMs with the high fidelity of implicit 3D GANs. Inspired by the impressive photorealism and expressive 3D representations of StyleSDF, our model uses a similar architecture but enforces the latent space to match the interpretable and physical parameters of the nonlinear 3D morphable model MOST-GAN. We demonstrate high-fidelity image synthesis and subsequent 3D manipulation with full control over the disentangled latent parameters.

关键词： Image quality Solid modeling Photorealism Three-dimensional displays Image synthesis Generative adversarial networks Rendering (computer graphics)

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：