检索结果-内蒙古大学图书馆

IEEE Workshop on Applications of Computer Vision (WACV)

作者： Jiaoyang Yin Bin Fan Chao Xu Tiejun Huang Boxin Shi State Key Lab of Multimedia Info. Processing School of Computer Science Peking University Nat'l Eng. Research Ctr. of Visual Technology School of Computer Science Peking University Nat'l Key Lab of General AI School of Intelligence Science and Technology Peking University

ISBN: (数字)9798331510831

ISBN: (纸本)9798331510848

As a bio-inspired vision sensor, the spiking camera has showcased remarkable capability in high-speed imaging with a sampling rate of 40,000 Hz. Reconstructing clear images from continuous spike streams, which is obtained by each photosensor continuously detecting photons and firing them asynchronously, has garnered significant attention. Despite promising results, existing spike-to-image reconstruction methods face challenges in balancing global receptive fields and efficient computation due to the inherent limitations of their backbones. Recently, due to powerful long-range modeling and linear complexity, the state space model (SSM) has emerged as a competitive alternative to CNNs and Transformers. In this paper, we propose a lightweight spike-to-image reconstruction network that harnesses Mamba as the backbone. Our approach sequentially executes three core modules: temporal information integration, spatial feature enhancement, and progressive image reconstruction. The former accumulates cues across diverse temporal windows to explore both long-term and short-term contexts. Subsequently, to model global dependencies while heightening local detail perception, we develop a multi-scale SSM block characterized by multi-scale multi-direction scanning, which effectively boosts spatial feature representations. Finally, intensity images are decoded progressively from the enhanced light-intensity features. Extensive experiments on both synthetic and real-captured data demonstrate that our approach achieves state-of-the-art performance, with only 10% of the network parameters and nearly two orders of magnitude less computational effort. The code will be available at https://***/interstellarH/Spk2ImgMamba.

关键词： Correlation Computational modeling reconstruction algorithms Vision sensors Transformer cores Streaming media Cameras Transformers image reconstruction Streams

来源：评论

学校读者我要写书评

暂无评论

image reconstruction from incomplete data VII

Image Reconstruction from Incomplete Data VII

引用

image reconstruction from incomplete data VII

ISBN: (纸本)9780819492173

The proceedings contain 26 papers. The topics discussed include: optical motion tracking to improve image quality in MRI of the brain;phase-space analysis of sparse signals and compressive sensing;compressive sampling methods for superresolution imaging;a Gibbs sampler for conductivity imaging and other inverse problems;imaging from scattered fields: limited data and degrees of freedom;joint time-frequency analysis of EEG signals based on a phase-space interpretation of the recording process;superresolved image reconstruction from incomplete data;multi-axial CT reconstruction from few view projections;using inverse fringe projection to speed up the detection of local and global geometry defects on free-form surfaces;support estimation for phase retrieval image reconstruction from sparse-aperture interferometry data;and adaptive binary material classification of an unknown object using polarimetric images degraded by atmospheric turbulence.

关键词：

来源：评论

学校读者我要写书评

暂无评论

ConDiSR: Contrastive Disentanglement and Style Regularization for Single Domain Generalization

ConDiSR: Contrastive Disentanglement and Style Regularizatio...

引用

IEEE Workshop on Applications of Computer Vision (WACV)

作者： Aleksandr Matsun Numan Saeed Fadillah Adamsyah Maani Mohammad Yaqub MBZUAI United Arab Emirates

ISBN: (数字)9798331510831

ISBN: (纸本)9798331510848

Medical data often exhibits distribution shifts, leading to performance degradation of deep learning models trained using standard supervised learning pipelines. Domain Generalization (DG) addresses this challenge, with Single-Domain Generalization (SDG) being notably relevant due to the privacy and logistical constraints often inherent in medical data. Existing disentanglement-based SDG methods heavily rely on structural information from segmentation masks, but classification labels do not offer similarly dense information. This work introduces a novel SDG method for medical image classification, utilizing channel-wise contrastive disentanglement. The method is further refined with reconstruction-based style regularization to ensure distinct style and structural feature representations are extracted. We evaluate our method on the complex tasks of multicenter histopathology image classification and Diabetic Retinopathy (DR) grading in fundus images, benchmarking it against state-of-the-art (SOTA) SDG baselines. Our results demonstrate that our method consistently outperforms the SOTA independently on the choice of the source domain while exhibiting greater performance stability. This study underscores the importance and challenges of exploring SDG frameworks for classification tasks. The code is publicly available at https://***/BioMedIA-MBZUAI/ConDiSR

关键词： image segmentation Histopathology Supervised learning Pipelines Feature extraction Stability analysis image reconstruction Standards Biomedical imaging image classification

来源：评论

学校读者我要写书评

暂无评论

Cross-Channel Unlabeled Sensing over a Union of Signal Subspaces

Cross-Channel Unlabeled Sensing over a Union of Signal Subsp...

引用

International conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Taulant Koka Manolis C. Tsakiris Benjamín Béjar Haro Michael Muma Robust Data Science Group Technische Universität Darmstadt Darmstadt Germany Academy of Mathematics and Systems Science Chinese Academy of Sciences Beijing China Swiss Data Science Center Paul Scherrer Institute Villigen Switzerland

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Cross-channel unlabeled sensing addresses the problem of recovering a multi-channel signal from measurements that were shuffled across channels. This work expands the cross-channel unlabeled sensing framework to signals that lie in a union of subspaces. The extension allows for handling more complex signal structures and broadens the framework to tasks like compressed sensing. These mismatches between samples and channels often arise in applications such as whole-brain calcium imaging of freely moving organisms or multi-target tracking. We improve over previous models by deriving tighter bounds on the required number of samples for unique reconstruction, while supporting more general signal types. The approach is validated through an application in whole-brain calcium imaging, where organism movements disrupt sample-to-neuron mappings. This demonstrates the utility of our framework in real-world settings with imprecise sample-channel associations, achieving accurate signal reconstruction.

关键词： Neuroscience Tracking Calcium Imaging Organisms Signal reconstruction Sensors Speech processing image reconstruction Compressed sensing

来源：评论

学校读者我要写书评

暂无评论

Using Conditional GAN for Facial Recognition Enhancement

Using Conditional GAN for Facial Recognition Enhancement

引用

International conference on ent Computing and Control Systems (ICICCS)

作者： T Pruthviraj Singh T Mahima Singh Deepak Kumar G Department of Cse-Cyber Security MITS MADANAPALLE Madanapalle India Department of Cse-Cyber Security AMRITA VISWA VIDYAPEETAM Chennai India

ISBN: (数字)9798331512088

ISBN: (纸本)9798331512095

Increasing dependence of technology presents issues for security agencies in identifying suspects, stemming from the inadequacies of eyewitness-based sketching techniques. Conventional methods fail to effectively convert sketches into lifelike facial representations, frequently resulting in misidentifications. This study advocates for the application of Conditional Generative Adversarial Networks (CGANs) to improve the realism and precision of suspect images produced from sketches. Conditional Generative Adversarial Networks utilize the deep learning methodologies to generate the high-quality photos by acquiring the feature representations from a dataset of facial images and their accompanying sketches. The main aim of this study is to connect the forensic sketching with digital image reconstruction, allowing the law enforcement authorities to produce more accurate and recognizable facial images based on witness testimonies. This work seeks to enhance the criminal investigations by enhancing the precision of sketch-to-image translation, hence the accelerating suspect identification and the minimizing erroneous identifications. The proposed model incorporates the data augmentation, a CGAN-based synthesis mechanism, and an intuitive interface for real-time image production. This study enhances forensic applications of AI, illustrating the significant influence of deep learning on law enforcement and public safety.

关键词： Training Law enforcement Predictive models Streaming media Generative adversarial networks data augmentation Prediction algorithms data models Real-time systems Generators

来源：评论

学校读者我要写书评

暂无评论

Seeing the Unseen in Micro-Video Popularity Prediction: Self-Correlation Retrieval for Missing Modality Generation 25

Seeing the Unseen in Micro-Video Popularity Prediction: Self...

引用

Proceedings of the 31st ACM SIGKDD conference on Knowledge Discovery and data Mining V.1

作者： Zhangtao Cheng Jian Lang Ting Zhong Fan Zhou University of Electronic Science and Technology of China Chengdu Sichuan China

ISBN: (纸本)9798400712456

Micro-video popularity prediction (MVPP) plays a crucial role in numerous real-world applications, including product marketing and recommendation systems. While existing methodologies predominantly assume complete modalities during multimodal learning, this assumption often fails to hold in practical scenarios due to various constraints, such as privacy concerns or data integrity issues. To address this limitation, we propose SCRAG, a novel Self-Correlation Retrieval-Augmented Generative framework designed to enhance missing-modality robustness in MVPP. SCRAG operates in a retrieval-guided generation manner that explores relevant knowledge to enhance the reconstruction of missing content, which consists of two primary components: (1) a self-correlation retriever and (2) a multimodal mixture-of-experts generator. It first acquires instances pertinent to the missing content through multimodal prompt alignment. Subsequently, the generator extracts contextual modal information from the retrieved context-rich instances. By learning the joint distribution of modalities, SCRAG effectively recovers missing content and addresses the modal heterogeneity challenge inherent in cross-modal generation approaches. Extensive experiments conducted on three real-world datasets demonstrate that SCRAG consistently outperforms state-of-the-art baselines, underscoring its effectiveness in handling incomplete modalities and improving the accuracy of micro-video popularity prediction.

关键词： incomplete multimodal learning

来源：评论

学校读者我要写书评

暂无评论

Enhancing Medical images with Generative Adversarial Networks (GANs): A Review of Techniques and Applications

Enhancing Medical Images with Generative Adversarial Network...

引用

Cognitive Computing in Engineering, Communications, Sciences and Biomedical Health Informatics (IC3ECSBHI), International conference on

作者： Harsh Shukla Saurabh Sharma Er. Deepika Dept. of Computer Science and Engg Chandigarh University Mohali India

ISBN: (数字)9798331518523

ISBN: (纸本)9798331518530

Generative Adversarial Networks (GANs) have demonstrated substantial utility in the domain of synthetic image generation for medical imaging. There are several obstacles to the growing need for high-quality photographs for illness detection, including expense, time, and radiation exposure. Generative adversarial networks have found applications in various image-to-image translation tasks such as virtual reconstruction of 7T MRIs from 3T images, enhancing low-quality resolution images into higher quality ones, and similar converting CT scans to MRIs. By creating multimodal datasets from a single modality, these methods greatly enhance data quality and eliminate the need for repeated scans. This paper examines several GAN models and how they are used in medical image processing, emphasizing how they may improve picture quality for more accurate diagnosis.

关键词： Training Accuracy Translation image synthesis Reviews Magnetic resonance imaging Superresolution Noise reduction Generative adversarial networks Medical diagnostic imaging

来源：评论

学校读者我要写书评

暂无评论

Cross-Modal Causal Representation Learning for Radiology Report Generation

引用

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society 2025年 PP卷 PP页

作者： Weixing Chen Yang Liu Ce Wang Jiarui Zhu Guanbin Li Cheng-Lin Liu Liang Lin School of Computer Science and Engineering Sun Yat-sen University Guangzhou China Guangdong Key Laboratory of Big Data Analysis and Processing Guangzhou China School of Science Sun Yat-sen University Shenzhen China Department of Health Technology and Informatics The Hong Kong Polytechnic University Hung Hom Hong Kong State Key Laboratory of Multimodal Artificial Intelligence Systems Institute of Automation Chinese Academy of Sciences Beijing China School of Artificial Intelligence University of Chinese Academy of Sciences Beijing China

Radiology Report Generation (RRG) is essential for computer-aided diagnosis and medication guidance, which can relieve the heavy burden of radiologists by automatically generating the corresponding radiology reports according to the given radiology image. However, generating accurate lesion descriptions remains challenging due to spurious correlations from visual-linguistic biases and inherent limitations of radiological imaging, such as low resolution and noise interference. To address these issues, we propose a two-stage framework named Cross-Modal Causal Representation Learning (CMCRL), consisting of the Radiological Cross-modal Alignment and reconstruction Enhanced (RadCARE) pre-training and the Visual-Linguistic Causal Intervention (VLCI) fine-tuning. In the pre-training stage, RadCARE introduces a degradation-aware masked image restoration strategy tailored for radiological images, which reconstructs high-resolution patches from low-resolution inputs to mitigate noise and detail loss. Combined with a multiway architecture and four adaptive training strategies (e.g., text postfix generation with degraded images and text prefixes), RadCARE establishes robust cross-modal correlations even with incomplete data. In the VLCI phase, we deploy causal front-door intervention through two modules: the Visual Deconfounding Module (VDM) disentangles local-global features without fine-grained annotations, while the Linguistic Deconfounding Module (LDM) eliminates context bias without external terminology databases. Experiments on IU-Xray and MIMIC-CXR show that our CMCRL pipeline significantly outperforms state-of-the-art methods, with ablation studies confirming the necessity of both stages. Code and models are available at.

关键词： Visualization Linguistics Radiology Heart Training Correlation Feature extraction Lesions Accuracy Lungs

来源：评论

学校读者我要写书评

暂无评论

An Effective Method for Checking Consistency in Univariate data-series with Multiple, incomplete and Inconsistent Entries

An Effective Method for Checking Consistency in Univariate D...

引用

International image Processing, Applications and Systems conference (IPAS)

作者： Sameesha Singh Maneesha Singh Railvision Europe Ltd Castle Donington UK Centre for Health Innovations MRIIRS Faridabad India

ISBN: (数字)9798331506520

ISBN: (纸本)9798331506537

Univariate time series is utilized in numerous scientific applications for a variety of purposes. They serve as the foundation for different statistical measurements and show how data evolves over time when viewed independently as real-time data series. Furthermore, binary data sets are frequently utilized in various fields, including health sciences, where each observation results in either a positive or negative outcome. The primary uses of time series analysis include: (a) Predicting future trends; (b) Determining characteristics like average, range, and standard deviation to comprehend the data's behavior; (c) Examining variations in means and variances among multiple time series using statistical methods like t-tests, F-tests, and Anova; and (d) Contrasting multiple time series to identify point-to-point discrepancies using different measures like root mean square error. This article suggests a straightforward yet successful method for comparing data from various time series that are not synchronized, have gaps in data, and require determining the error of the best match to ensure it falls within an acceptable range. We utilize this method for a challenging railway engineering issue to verify the consistency of laser and inertial sensor measurements within a specified error range.

关键词： Railway engineering Time series analysis Measurement uncertainty Measurement by laser beam Market research Time measurement Real-time systems Synchronization Root mean square Standards

来源：评论

学校读者我要写书评

暂无评论

Rapid three-dimensional seed reconstruction from incomplete data sets for image guided prostate brachytherapy

Rapid three-dimensional seed reconstruction from incomplete ...

引用

作者： Narayanan, Sreeram University of Washington

学位级别：Ph.D.

Prostate cancer is one of the most common forms of cancer among men. Treatment of prostate cancer involves implanting radioactive seeds into the prostate, a procedure known as prostate brachytherapy. Brachytherapy involves implanting seeds precisely into locations as defined by the plan. Any deviation from the plan could lead to lower cure rates. Dosimetry calculation involves computing the 3D locations of the seeds. Two major challenges for performing dosimetry in the operating room (intraoperative dosimetry) are (a) speed of computation and (b) undetected seeds on any of the three 2D images. We propose an epipolar geometry based reconstruction algorithm that is computationally efficient and can handle incomplete data sets. The algorithm divides the images into bands for speed of matching and reconstruction and uses a duplicate and match methodology incase any of the bands have undetected seeds. We test the algorithm on 205 patient cases. In order to obtain a complete picture of the doses the prostate boundary needs to be visualized along with the seeds. Prostate boundary is not visualized on fluoroscopy, the imaging modality used for seed reconstruction. Ultrasound is a modality that has a good prostate visualization but poor seed visualization. The two imaging modalities can be used for their strengths and data fusion can be performed to get the complete picture. We provide three fusion techniques (a) dummy probe based fusion; (b) embedded marker based fusion and (c) fusion using seeds themselves. The algorithm was tested using a few phantoms and also on patient cases.

关键词：

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：