检索结果-内蒙古大学图书馆

FADE: A Task-Agnostic Upsampling Operator for Encoder-Decoder Architectures

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Lu, Hao Liu, Wenze Fu, Hongtao Cao, Zhiguo The Key Laboratory of Image Processing and Intelligent Control Ministry of Education School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China

The goal of this work is to develop a task-agnostic feature upsampling operator for dense prediction where the operator is required to facilitate not only region-sensitive tasks like semantic segmentation but also detail-sensitive tasks such as image matting. Prior upsampling operators often can work well in either type of the tasks, but not both. We argue that task-agnostic upsampling should dynamically trade off between semantic preservation and detail delineation, instead of having a bias between the two properties. In this paper, we present FADE, a novel, plug-and-play, lightweight, and task-agnostic upsampling operator by fusing the assets of decoder and encoder features at three levels: i) considering both the encoder and decoder feature in upsampling kernel generation;ii) controlling the per-point contribution of the encoder/decoder feature in upsampling kernels with an efficient semi-shift convolutional operator;and iii) enabling the selective pass of encoder features with a decoder-dependent gating mechanism for compensating details. To improve the practicality of FADE, we additionally study parameter- and memory-efficient implementations of semi-shift convolution. We analyze the upsampling behavior of FADE on toy data and show through large-scale experiments that FADE is task-agnostic with consistent performance improvement on a number of dense prediction tasks with little extra cost. For the first time, we demonstrate robust feature upsampling on both region- and detail-sensitive tasks successfully. Code is made available at: https://***/poppinace/fade Copyright © 2024, The Authors. All rights reserved.

关键词： Object detection

Fast image Matching Based on Multi-core DSP

学校读者我要写书评

暂无评论

Fast Image Matching Based on Multi-core DSP

International Conference on intelligent Computation and Bio-Medical Instrumentation (ICBMI)

作者： Fumin Guo Mingyue Ding Xuming Zhang School of Life Science and Technology Key Laboratory of Image Processing and Intelligent Control of Education Ministry of China Huazhong University of Science and Technology Wuhan China

Traditional image matching algorithm based on gray correlation provides accurate results but it is time-consuming because of large amount of calculation. An improved gray correlation based image matching algorithm based on multi-core DSP is proposed to speed up the matching velocity with the acceptable margin of errors. With the development of portable embedded image processer, especially the multi-core DSPs for parallel computing to speed up the process, these image matching algorithms need to be transplanted to these embedded system. Experiments based on CPU in the form of Visual C++6.0 application program and multi-core Digital Signal Processor(DSP) verify the effectiveness of the algorithm, making it applicable to embedded image processing system with a multi-core DSP.

关键词： Digital signal processing Correlation image matching Signal processing algorithms Algorithm design and analysis Multicore processing Feature extraction

Pseudo-Labeling by Multi-Policy Viewfinder Network for image Cropping

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Pan, Zhiyu Wang, Kewei Wu, Yizheng Xiao, Liwen Cui, Jiahao Wang, Zhicheng Cao, Zhiguo Key Laboratory of Image Processing and Intelligent Control Ministry of Education School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China

Automatic image cropping models predict reframing boxes to enhance image aesthetics. Yet, the scarcity of labeled data hinders the progress of this task. To overcome this limitation, we explore the possibility of utilizing both labeled and unlabeled data together to expand the scale of training data for image cropping models. This idea can be implemented in a pseudo-labeling way: producing pseudo labels for unlabeled data by a teacher model and training a student model with these pseudo labels. However, the student may learn from teacher's mistakes. To address this issue, we propose the multi-policy viewfinder network (MPV-Net) that offers diverse refining policies to rectify the mistakes in original pseudo labels from the teacher. The most reliable policy is selected to generate trusted pseudo labels. The reliability of policies is evaluated via the robustness against box jittering. The efficacy of our method can be evaluated by the improvement compared to the supervised baseline which only uses labeled data. Notably, our MPV-Net outperforms off-the-shelf pseudo-labeling methods, yielding the most substantial improvement over the supervised baseline. Furthermore, our approach achieves state-of-the-art results on both the FCDB and FLMS datasets, signifying the superiority of our approach. Copyright © 2024, The Authors. All rights reserved.

关键词： Distillation

Thermal Safety Analysis and Dynamic control Strategy for Mode Switching Process of Reversible Solid Oxide Cell System

学校读者我要写书评

暂无评论

Thermal Safety Analysis and Dynamic Control Strategy for Mod...

Chinese control Conference (CCC)

作者： Wei Zhao Guoqiang Liu Zexin Li Xi Li Key Laboratory of Image Processing and Intelligent Control of Education Ministry School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan Hubei China

Reversible solid oxide cells (rSOC) can operate in both electrolysis mode and in fuel cell mode with high efficiency and reduced cost using the same device for both functions. When used in dynamic operation with intermittent electrical power sources, rSOC system switches from the fuel cell (SOFC) to electrolyzer (SOEC) and vice versa depending on load and grid peculiarities. This can lead to temperature profiles within the stack that can potentially lead to the failure of the stack and eventually the system. In the present work, a system-level dynamic model of rSOC is established and validated against experimental data. Subsequently, detailed dynamic thermal behavior of stack during the switching between the two modes is analyzed, and a temperature management controller based on Model Predict control method (MPC) is proposed. Gas flow and temperature at the stack inlet are controlled by air flow and bypass valves, avoiding problems associated with temperature overshoot during transient operation. The simulation results show that the temperature controller has the ability to follow fast thermal changes while maintaining thermal safety, which indicates the competitiveness of the controller.

关键词：

Knowledge-Data Fusion Based Source-Free Semi-Supervised Domain Adaptation for Seizure Subtype Classification

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Peng, Ruimin An, Jiayu Wu, Dongrui Key Laboratory of the Ministry of Education for Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China

Electroencephalogram (EEG)-based seizure subtype classification enhances clinical diagnosis efficiency. Source-free semi-supervised domain adaptation (SF-SSDA), which transfers a pre-trained model to a new dataset with no source data and limited labeled target data, can be used for privacy-preserving seizure subtype classification. This paper considers two challenges in SF-SSDA for EEG-based seizure subtype classification: 1) How to effectively fuse both raw EEG data and expert knowledge in classifier design? 2) How to align the source and target domain distributions for SF-SSDA? We propose a Knowledge-Data Fusion based SF-SSDA approach, KDF-MutualSHOT, for EEG-based seizure subtype classification. In source model training, KDF uses Jensen-Shannon Divergence to facilitate mutual learning between a feature-driven Decision Tree-based model and a data-driven Transformer-based model. To adapt KDF to a new target dataset, an SF-SSDA algorithm, MutualSHOT, is developed, which features a consistency-based pseudo-label selection strategy. Experiments on the public TUSZ and CHSZ datasets demonstrated that KDF-MutualSHOT outperformed other supervised and source-free domain adaptation approaches in cross-subject seizure subtype classification. © 2024, CC BY.

关键词： Self-supervised learning

Geometric Active Contour Model and its Application to Carotid Plaque Detection

学校读者我要写书评

暂无评论

Geometric Active Contour Model and its Application to Caroti...

International Conference on intelligent Computation and Bio-Medical Instrumentation (ICBMI)

作者： Jieyu Cheng Mingyue Ding Xuming Zhang Life School of Science and Technology Key Laboratory of Image Processing and Intelligent Control of Education Ministry of China Huazhong University of Science and Technology Wuhan China

This paper gives an overview of three different geometric active contour models with the focus on the application to carotid plaque detection from cross-sectional ultrasound images of the carotid artery. On one hand, basic principles of these geometric active contour models are presented. On the other hand, performance of these models are tested on 16 images and compared with the manual delineations.

关键词： Active contours image segmentation Computational modeling Manuals Level set Ultrasonic imaging Solid modeling

Method of boundary extraction based on schrödinger equation 21

学校读者我要写书评

暂无评论

Method of boundary extraction based on schrödinger equation

2008 21st ISPRS International Congress for Photogrammetry and Remote Sensing

作者： Lou, Liantang Zhan, Xin Fu, Zhongliang Ding, Mingyue School of Remote Sensing and Information Engineering Wuhan University Wuhan430072 China School of Science Wuhan Institute of Technology Wuhan430073 China Institute for Pattern Recognition and Artificial Intelligence Key Laboratory of Education Ministry for Image Processing and Intelligent Control Huazhong University of Science and Technology Wuhan430074 China

To overcome the main drawbacks of global minimal for active contour models (L. D. Cohen and Ron Kimmel) that the contour is only extracted partially for low SNR images, Method of boundary extraction based on Schrödinger Equation is proposed. Our Method is based on computing the numerical solutions of initial value problem for second order nonlinear Schrödinger equation by using discrete Fourier Transformation. Schrödinger transformation of image is first given. We compute the probability P(b,a) that a particle moves from a point a to another point b according to I-Type Schrödinger transformation of image and obtain boundary of object by using quantum contour model.

关键词： Initial value problems

When Epipolar Constraint Meets Non-local Operators in Multi-View Stereo

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Liu, Tianqi Ye, Xinyi Zhao, Weiyue Pan, Zhiyu Shi, Min Cao, Zhiguo Key Laboratory of Image Processing and Intelligent Control Ministry of Education School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China

Learning-based multi-view stereo (MVS) method heavily relies on feature matching, which requires distinctive and descriptive representations. An effective solution is to apply non-local feature aggregation, e.g., Transformer. Albeit useful, these techniques introduce heavy computation overheads for MVS. Each pixel densely attends to the whole image. In contrast, we propose to constrain non-local feature augmentation within a pair of lines: each point only attends the corresponding pair of epipolar lines. Our idea takes inspiration from the classic epipolar geometry, which shows that one point with different depth hypotheses will be projected to the epipolar line on the other view. This constraint reduces the 2D search space into the epipolar line in stereo matching. Similarly, this suggests that the matching of MVS is to distinguish a series of points lying on the same line. Inspired by this point-to-line search, we devise a line-to-point non-local augmentation strategy. We first devise an optimized searching algorithm to split the 2D feature maps into epipolar line pairs. Then, an Epipolar Transformer (ET) performs non-local feature augmentation among epipolar line pairs. We incorporate the ET into a learning-based MVS baseline, named ET-MVSNet. ET-MVSNet achieves state-of-the-art reconstruction performance on both the DTU and Tanks-and-Temples benchmark with high efficiency. Code is available at https://***/TQTQliu/ET-MVSNet. Copyright © 2023, The Authors. All rights reserved.

关键词： Benchmarking

A3E: Aligned and Augmented Adversarial Ensemble for Accurate, Robust and Privacy-Preserving EEG Decoding

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Chen, Xiaoqing Jia, Tianwang Wu, Dongrui Key Laboratory of Image Processing and Intelligent Control Ministry of Education School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China

An electroencephalogram (EEG) based brain-computer interface (BCI) enables direct communication between the brain and external devices. However, EEG-based BCIs face at least three major challenges in real-world applications: data scarcity and individual differences, adversarial vulnerability, and data privacy. While previous studies have addressed one or two of these issues, simultaneous accommodation of all three challenges remains challenging and unexplored. This paper fills this gap, by proposing an Aligned and Augmented Adversarial Ensemble (A3E) algorithm and integrating it into three privacy protection scenarios (centralized source-free transfer, federated source-free transfer, and source data perturbation), achieving simultaneously accurate decoding, adversarial robustness, and privacy protection of EEG-based BCIs. Experiments on three public EEG datasets demonstrated that our proposed approach outperformed over 10 classic and state-of-the-art approaches in both accuracy and robustness in all three privacy-preserving scenarios, even outperforming state-of-the-art transfer learning approaches that do not consider privacy protection at all. This is the first time that three major challenges in EEG-based BCIs can be addressed simultaneously, significantly improving the practicalness of EEG decoding in real-world BCIs. © 2024, CC BY.

关键词： Transfer learning