检索结果-内蒙古大学图书馆

Reflash Dropout in Image Super-Resolution

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Kong, Xiangtao Liu, Xina Gu, Jinjin Qiao, Yu Dong, Chao ShenZhen Key Lab of Computer Vision and Pattern Recognition SIAT-SenseTime Joint Lab. Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences China University of Chinese Academy of Sciences China The University of Sydney Australia Shanghai AI Laboratory Shanghai China

Dropout is designed to relieve the overfitting problem in high-level vision tasks but is rarely applied in low-level vision tasks, like image super-resolution (SR). As a classic regression problem, SR exhibits a different behaviour as high-level tasks and is sensitive to the dropout operation. However, in this paper, we show that appropriate usage of dropout benefits SR networks and improves the generalization ability. Specifically, dropout is better embedded at the end of the network and is significantly helpful for the multi-degradation settings. This discovery breaks our common sense and inspires us to explore its working mechanism. We further use two analysis tools – one is from a recent network interpretation work, and the other is specially designed for this task. The analysis results provide side proofs to our experimental findings and show us a new perspective to understand SR networks. Copyright © 2021, The Authors. All rights reserved.

关键词： Optical resolving power

Low-Resolution Action recognition for Tiny Actions Challenge

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Chen, Boyu Qiao, Yu Wang, Yali ShenZhen Key Lab of Computer Vision and Pattern Recognition Shenzhen Institute of Advanced Technology Chinese Academy of Sciences China University of Chinese Academy of Sciences China Shanghai AI Laboratory Shanghai China SIAT Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society China

Tiny Actions Challenge focuses on understanding human activities in real-world surveillance. Basically, there are two main difficulties for activity recognition in this scenario. First, human activities are often recorded at a distance, and appear in a small resolution without much discriminative clue. Second, these activities are naturally distributed in a long-tailed way. It is hard to alleviate data bias for such heavy category imbalance. To tackle these problems, we propose a comprehensive recognition solution in this paper. First, we train video backbones with data balance, in order to alleviate overfitting in the challenge benchmark. Second, we design a dual-resolution distillation framework, which can effectively guide low-resolution action recognition by super-resolution knowledge. Finally, we apply model ensemble with post-processing, which can further boost performance on the long-tailed categories. Our solution ranks Top-1 on the leaderboard. Copyright © 2022, The Authors. All rights reserved.

关键词： Distillation

PPT Fusion: Pyramid Patch Transformer for a Case Study in Image Fusion

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Fu, Yu Xu, Tianyang Wu, Xiao-Jun Kittler, Josef Jiangsu Provincial Engineering Laboratory of Pattern Recognition and Computational Intelligence School of Artificial Intelligence and Computer Science Jiangnan University Wuxi 214122 China Centre for Vision Speech and Signal Processing University of Surrey GuildfordGU2 7XH United Kingdom

The Transformer architecture has witnessed a rapid development in recent years, outperforming the CNN architectures in many computer vision tasks, as exemplified by the vision Transformers (ViT) for image classification. However, existing visual transformer models aim to extract semantic information for high-level tasks, such as classification and detection. These methods ignore the importance of the spatial resolution of the input image, thus sacrificing the local correlation information of neighboring pixels. In this paper, we propose a Patch Pyramid Transformer(PPT) to effectively address the above issues. Specifically, we first design a Patch Transformer to transform the image into a sequence of patches, where transformer encoding is performed for each patch to extract local representations. In addition, we construct a Pyramid Transformer to effectively extract the non-local information from the entire image. After obtaining a set of multi-scale, multi-dimensional, and multi-angle features of the original image, we design the image reconstruction network to ensure that the features can be reconstructed into the original input. To validate the effectiveness, we apply the proposed Patch Pyramid Transformer to image fusion tasks. The experimental results demonstrate its superior performance, compared to the state-of-the-art fusion approaches, achieving the best results on several evaluation indicators. Thanks to the underlying representational capacity of the PPT network, it can directly be applied to different image fusion tasks without redesigning or retraining the network. Copyright © 2021, The Authors. All rights reserved.

关键词： Deep learning

Exploring Fusion Strategies for Accurate RGBT Visual Object Tracking

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Tang, Zhangyong Xu, Tianyang Li, Hui Wu, Xiao-Jun Zhu, XueFeng Kittler, Josef Jiangsu Provincial Engineering Laboratory of Pattern Recognition and Computational Intelligence School of Artificial Intelligence and Computer Science Jiangnan University Wuxi214122 China The Center for Vision Speech and Signal Processing University of Surrey GuildfordGU2 7XH United Kingdom

We address the problem of multi-modal object tracking in video and explore various options of fusing the complementary information conveyed by the visible (RGB) and thermal infrared (TIR) modalities including pixel-level, feature-level and decision-level fusion. Specifically, different from the existing methods, paradigm of image fusion task is heeded for fusion at pixel level. Feature-level fusion is fulfilled by attention mechanism with channels excited optionally. Besides, at decision level, a novel fusion strategy is put forward since an effortless averaging configuration has shown the superiority. The effectiveness of the proposed decision-level fusion strategy owes to a number of innovative contributions, including a dynamic weighting of the RGB and TIR contributions and a linear template update operation. A variant of which produced the winning tracker at the Visual Object Tracking Challenge 2020 (VOT-RGBT2020). The concurrent exploration of innovative pixel- and feature-level fusion strategies highlights the advantages of the proposed decision-level fusion method. Extensive experimental results on three challenging datasets, i.e., GTOT, VOT-RGBT2019, and VOT-RGBT2020, demonstrate the effectiveness and robustness of the proposed method, compared to the state-of-the-art approaches. Code will be shared at https://***/Zhangyong-Tang/DFAT. © 2022, CC BY.

关键词： Pixels

RFN-Nest: An end-to-end residual fusion network for infrared and visible images

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Li, Hui Wu, Xiao-Jun Kittler, Josef Jiangsu Provincial Engineering Laboratory of Pattern Recognition and Computational Intelligence School of Artificial Intelligence and Computer Science Jiangnan University Wuxi214122 China The Center for Vision Speech and Signal Processing University of Surrey GuildfordGU2 7XH United Kingdom

In the image fusion field, the design of deep learning-based fusion methods is far from routine. It is invariably fusion-task specific and requires a careful consideration. The most difficult part of the design is to choose an appropriate strategy to generate the fused image for a specific task in hand. Thus, devising learnable fusion strategy is a very challenging problem in the community of image fusion. To address this problem, a novel end-to-end fusion network architecture (RFN-Nest) is developed for infrared and visible image fusion. We propose a residual fusion network (RFN) which is based on a residual architecture to replace the traditional fusion approach. A novel detail-preserving loss function, and a feature enhancing loss function are proposed to train RFN. The fusion model learning is accomplished by a novel two-stage training strategy. In the first stage, we train an auto-encoder based on an innovative nest connection (Nest) concept. Next, the RFN is trained using the proposed loss functions. The experimental results on public domain data sets show that, compared with the existing methods, our end-to-end fusion network delivers a better performance than the state-of-the-art methods in both subjective and objective evaluation. The code of our fusion method is available at https://***/hli1221/imagefusion-rfn-nest. © 2021, CC BY-NC-ND.

关键词： Image fusion

EDEN: Deep feature distribution pooling for Saimaa ringed seals pattern matching

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Chelak, Ilia Nepovinnykh, Ekaterina Eerola, Tuomas Kälviäinen, Heikki Belykh, Igor Peter the Great St. Petersburg Polytechnic University Saint Petersburg Russia Lappeenranta-Lahti University of Technology LUT School of Engineering Science Department of Computational Engineering Computer Vision and Pattern Recognition Laboratory P.O.Box 20 Lappeenranta53850 Finland

In this paper, pelage pattern matching is considered to solve the individual re-identification of the Saimaa ringed seals. Animal re-identification together with the access to large amount of image material through camera traps and crowd-sourcing provide novel possibilities for animal monitoring and conservation. We propose a novel feature pooling approach that allow aggregating the local pattern features to get a fixed size embedding vector that incorporate global features by taking into account the spatial distribution of features. This is obtained by eigen decomposition of covariances computed for probability mass functions representing feature maps. Embedding vectors can then be used to find the best match in the database of known individuals allowing animal re-identification. The results show that the proposed pooling method outperforms the existing methods on the challenging Saimaa ringed seal image data. © 2021, CC BY-NC-ND.

关键词： pattern matching

rPPG-MAE: Self-supervised Pre-training with Masked Autoencoders for Remote Physiological Measurement

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Liu, Xin Zhang, Yuting Yu, Zitong Lu, Hao Yue, Huanjing Yang, Jingyu School of Electrical and Information Engineering Tianjin University Tianjin300072 China Computer Vision and Pattern Recognition Laboratory School of Engineering Science Lappeenranta-Lahti University of Technology LUT Lappeenranta53850 Finland Great Bay University Dongguan523000 China Guangzhou511453 China

Remote photoplethysmography (rPPG) is an important technique for perceiving human vital signs, which has received extensive attention. For a long time, researchers have focused on supervised methods that rely on large amounts of labeled data. These methods are limited by the requirement for large amounts of data and the difficulty of acquiring ground truth physiological signals. To address these issues, several self-supervised methods based on contrastive learning have been proposed. However, they focus on the contrastive learning between samples, which neglect the inherent self-similar prior in physiological signals and seem to have a limited ability to cope with noisy. In this paper, a linear self-supervised reconstruction task was designed for extracting the inherent self-similar prior in physiological signals. Besides, a specific noise-insensitive strategy was explored for reducing the interference of motion and illumination. The proposed framework in this paper, namely rPPG-MAE, demonstrates excellent performance even on the challenging VIPL-HR dataset. We also evaluate the proposed method on two public datasets, namely PURE and UBFC-rPPG. The results show that our method not only outperforms existing self-supervised methods but also exceeds the state-of-the-art (SOTA) supervised methods. One important observation is that the quality of the dataset seems more important than the size in self-supervised pre-training of rPPG. The source code is released at https://***/linuxsino/rPPG-MAE. Copyright © 2023, The Authors. All rights reserved.

关键词： Supervised learning

Dynamic hand gesture early recognition based on Hidden Semi-Markov Models

学校读者我要写书评

暂无评论

Dynamic hand gesture early recognition based on Hidden Semi-...

IEEE International Conference on Robotics and Biomimetics

作者： Qianqian Wang Yuanrong Xu Yen-Lun Chen Yong Wang Xinyu Wu University of Science and Technology of China Shenzhen Key Lab for Computer Vision and Pattern Recognition Chinese Academy of Sciences Dept. Mechanical and Automation Engineering The Chinese University of Hong Kong. Guangdong Provincial Key Laboratory of Robotics and Intelligent System Chinese Academy of Sciences

ISBN: (纸本)9781479973989

Real-time performance can be greatly improved, if the early recognition is implemented. In this paper, a dynamic hand gesture early recognition system is proposed. The system can recognize the gesture before it is completed. Our method is based on the Hidden Semi-Markov Models. Three-dimensional information of the gesture trajectory collected by leapmotion is the main data we used. Experiments on the dataset which we established demonstrate the effectiveness of our method.

关键词： Hidden Markov models Gesture recognition Training Conferences Trajectory Arrays

Comparison of appearance-based and geometry-based bubble detectors

学校读者我要写书评

暂无评论

Lecture Notes in Computer Science (including subseries Lectu...

Lecture Notes in computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 2014年 8671卷 610-617页

作者： Strokina, Nataliya Juránek, Roman Eerola, Tuomas Lensu, Lasse Zemčik, Pavel Kälviäinen, Heikki Tampere University of Technology Department of Signal Processing P.O. Box 527 Tampere33101 Finland Brno University of Technology Department of Computer Graphics and Multimedia Brno Czech Republic Lappeenranta University of Technology Machine Vision and Pattern Recognition Laboratory P.O. Box 20 Lappeenranta53851 Finland

Bubble detection is a complicated tasks since varying lighting conditions changes considerably the appearance of bubbles in liquid. The two common techniques to detect circular objects such as bubbles, the geometry-based and appearance-based approaches, have their advantages and weaknesses. The geometry-based methods often fail to detect small blob-like bubbles that do not match the used geometrical model, and appearance-based approaches are vulnerable to appearance changes caused by, e.g., illumination. In this paper, we compare a geometry-based concentric circular arrangements (CCA) and appearance-based sliding window methods as well as their combinations in terms of bubble detection, gas volume computation, and size distribution estimation. The best bubble detection performance was achieved with the sliding window method whereas the most precise volume estimate was produced by the CCA method. The combination of the two approaches gave only a minor advantage compared to the base methods. © Springer International Publishing Switzerland 2014.

关键词： Geometry