检索结果-内蒙古大学图书馆

7th International Conference on Medical Imaging with Deep Learning, MIDL 2024

作者： Buess, Lukas Stollenga, Marijn F. Schinz, David Wiestler, Benedikt Kirschke, Jan S. Maier, Andreas Navab, Nassir Keicher, Matthias Technical University of Munich Germany Klinikum Rechts der Isar Technical University of Munich Germany Pattern Recognition Lab. Friedrich-Alexander-Universität Erlangen-Nürnberg Germany ImFusion GmbH Munich Germany

Early and accurate diagnosis of vertebral body anomalies is crucial for effectively treating spinal disorders, but the manual interpretation of CT scans can be time-consuming and error-prone. While deep learning has shown promise in automating vertebral fracture detection, improving the interpretability of existing methods is crucial for building trust and ensuring reliable clinical application. Vision Transformers (ViTs) offer inherent interpretability through attention visualizations but are limited in their application to 3D medical images due to reliance on 2D image pretraining. To address this challenge, we propose a novel approach combining the benefits of transfer learning from video-pretrained models and domain adaptation with self-supervised pretraining on a task-specific but unlab.led dataset. Compared to naïve transfer learning from Video MAE, our method shows improved downstream task performance by 8.3 in F1 and a training speedup of factor 2. This closes the gap between videos and medical images, allowing a ViT to learn relevant anatomical features while adapting to the task domain. We demonstrate that our framework enables ViTs to effectively detect vertebral fractures in a low data regime, outperforming CNN-based state-of-the-art methods while providing inherent interpretability. Our task adaptation approach and dataset not only improve the performance of our proposed method but also enhance existing self-supervised pretraining approaches, highlighting the benefits of task-specific self-supervised pretraining for domain adaptation. The code is publicly availab.e at https://***/lbuess/Video-CT_MAE. © 2024 CC-BY 4.0, L. Buess et al.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

On the Use of a Foundation Acoustic Model to Identify Highly Relevant Phonetic Information of Parkinson’s Speech 11th

On the Use of a Foundation Acoustic Model to Identify Hi...

引用

11th Workshop on Engineering Applications, WEA 2024

作者： Escobar-Grisales, D. Ríos-Urrego, C.D. Orozco-Arroyave, J.R. GITA Lab Faculty of Engineering University of Antioquia UdeA Medellín Colombia Pattern Recognition Lab. Friedrich-Alexander Universität Erlangen-Nürnberg Erlangen Germany

ISBN: (纸本)9783031745942

Parkinson’s disease (PD) is a neurological condition that produces several speech deficits, typically known as hypokinetic dysarthria, affecting the production of different phonemes and resulting in an impaired speech communication. This work presents a detailed investigation based on the wav2vec 2.0 foundational model specifically tuned to perform the automatic discrimination between PD and healthy control (HC) subjects. The investigation showed that, instead of considering the complete wav2vec 2.0 architecture with 12 layers, the five layer is enough to find a model suitable to obtain good classification accuracies. Besides, this work presents a framework where frame-wise classification results are considered, enabling a detailed analysis regarding which phonemes and phonological classes are more accurate for performing the classification. All experiments are evaluated in an external and independent test set, therefore given the good results found in this work, which motivates us to continue working in this direction. For future work, we plan to modify the method to perform the time-stamp lab.ling to model co-articulation information in speech produced by PD patients. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Neurodegenerative diseases

来源：评论

学校读者我要写书评

暂无评论

Analysis of Argument Structure Constructions in a Deep Recurrent Language Model

arXiv

引用

arXiv 2024年

作者： Ramezani, Pegah Schilling, Achim Krauss, Patrick Department of English and American Studies University Erlangen-Nuremberg Germany Neuroscience Lab. University Hospital Erlangen Germany Cognitive Computational Neuroscience Group Pattern Recognition Lab. University Erlangen-Nuremberg Germany

Understanding how language and linguistic constructions are processed in the brain is a fundamental question in cognitive computational neuroscience. In this study, we explore the representation and processing of Argument Structure Constructions (ASCs) in a recurrent neural language model. We trained a Long Short-Term Memory (LSTM) network on a custom-made dataset consisting of 2000 sentences, generated using GPT-4, representing four distinct ASCs: transitive, ditransitive, caused-motion, and resultative constructions. We analyzed the internal activations of the LSTM model’s hidden layers using Multidimensional Scaling (MDS) and t-Distributed Stochastic Neighbor Embedding (t-SNE) to visualize the sentence representations. The Generalized Discrimination Value (GDV) was calculated to quantify the degree of clustering within these representations. Our results show that sentence representations form distinct clusters corresponding to the four ASCs across all hidden layers, with the most pronounced clustering observed in the last hidden layer before the output layer. This indicates that even a relatively simple, brain-constrained recurrent neural network can effectively differentiate between various construction types. These findings are consistent with previous studies demonstrating the emergence of word class and syntax rule representations in recurrent language models trained on next word prediction tasks. In future work, we aim to validate these results using larger language models and compare them with neuroimaging data obtained during continuous speech perception. This study highlights the potential of recurrent neural language models to mirror linguistic processing in the human brain, providing valuable insights into the computational and neural mechanisms underlying language understanding. © 2024, CC BY-NC-ND.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Cross-Layer Feature based Multi-Granularity Visual Classification

Cross-Layer Feature based Multi-Granularity Visual Classific...

引用

IEEE Visual Communications and Image Processing (VCIP)

作者： Junhan Chen Dongliang Chang Jiyang Xie Ruoyi Du Zhanyu Ma Pattern Recognition and Intelligent System Lab. Beijing University of Posts and Telecommunications Beijing China

ISBN: (纸本)9781665475938

In contrast to traditional fine-grained visual clas-sification, multi-granularity visual classification is no longer limited to identifying the different sub-classes belonging to the same super-class (e.g., bird species, cars, and aircraft models). Instead, it gives a sequence of lab.ls from coarse to fine (e.g., Passeriformes → Corvidae → Fish Crow), which is more convenient in practice. The key to solving this task is how to use the relationships between the different levels of lab.ls to learn feature representations that contain different levels of granularity. Interestingly, the feature pyramid structure naturally implies different granularity of feature representation, with the shallow layers representing coarse-grained features and the deep layers representing fine-grained features. Therefore, in this paper, we exploit this property of the feature pyramid structure to decouple features and obtain feature representations corre-sponding to different granularities. Specifically, we use shallow features for coarse-grained classification and deep features for fine-grained classification. In addition, to enable fine-grained features to enhance the coarse-grained classification, we propose a feature reinforcement module based on the feature pyramid structure, where deep features are first upsampled and then combined with shallow features to make decisions. Experimental results on three widely used fine-grained image classification datasets such as CUB-200-2011, Stanford Cars, and FGVC-Aircraft validate the method's effectiveness. Code availab.e at https://***/PRIS-CV/CGVC.

关键词： Visualization Cross layer design Codes Visual communication Image processing Fish Birds

来源：评论

学校读者我要写书评

暂无评论

PickDet: A Detection Framework for Aerial-view Scene

PickDet: A Detection Framework for Aerial-view Scene

引用

IEEE Visual Communications and Image Processing (VCIP)

作者： Cheng Lyu Xiao Deng Shizun Wang Ming Wu Chuang Zhang Pattern Recognition and Intelligent Systems Lab. Beijing University of Posts and Telecommunications Beijing China

ISBN: (纸本)9781665475938

Detecting objects in the aerial-view scene is challenging for the objects usually have small scales relative to the image, making it hard to achieve high accuracy in full-image detection. Slice detection tries to overcome this by cutting the full image into slices before detecting them, but objects are sparsely distributed and usually clustered in local areas, a large number of background areas without objects can be ignored to improve detection efficiency. In this paper, we present PickDet, a framework for efficient and effective object detection in the aerial-view scene, which only chooses slices containing objects to conduct detection. The key components of PickDet include a lightweight convolutional network (PickNet), a screening strategy (SoftPick), and fine-tuned detectors. Given slices of aerial-view images, PickNet first outputs the probability of object existence. Then SoftPick conducts a double-threshold screening strategy to pick the slices which contain objects. Finally, all picked slices are fed into the detector in parallel and full-image detection is used as an auxiliary mean. Compared with previous methods, PickDet achieves higher accuracy and more efficiency in the aerial-view scene. We evaluate PickDet on Visdrone and Oiltank datasets, experiments show that PickDet can result in up to 28.0% AP improvement compared to full-image detection, and can result in up to 2.9% AP increase and up to 5 times inference speedup compared to slice detection.

关键词： Fault tolerance Visual communication Image processing Fault tolerant systems Detectors Object detection

来源：评论

学校读者我要写书评

暂无评论

Assessment of Handwriting in Patients With Parkinson’s Disease Using Non-Intrusive Tasks

Assessment of Handwriting in Patients With Parkinson’s Dise...

引用

IEEE International Symposium on Biomedical Imaging

作者： J.D. Gallo-Aristizabal D. Escobar-Grisales C.D. Ríos-Urrego P. A. Pérez-Toro E. Nöth A. Maier J. R. Orozco-Arroyave GITA Lab Faculty of Engineering University of Antioquia UdeA Medellín Colombia Pattern Recognition Lab. Friedrich-Alexander Universität Erlangen-Nürnberg Erlangen Germany

This study presents two approaches for modeling the handwriting of Parkinson’s Disease (PD) patients and Healthy Control (HC) subjects. One approach is based on digit-embeddings generated from a CNN architecture pre-trained with information from the MNIST corpus. The second approach consists of the computation of statistical functionals of dynamics signal collected with the digital tablet, namely azimuth, pressure, altitude, and vertical distance. The experiments are based on writing the ten digits (from 0 to 9), which is a task commonly performed in daily life activities, making this approach closer to a non-intrusive evaluation. According to the results, the accuracy of the classification between PD patients vs. HC improved from 71.8% to 74.5% when information from images is combined with the functionals of the vertical distance signal.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Intelligent Model for Predicting the Quality of Services Violation

引用

Computers, Materials & Continua 2022年第5期71卷 3607-3619页

作者： Muhammad Adnan Khan Asma Kanwal Sagheer Abbas Faheem Khan T.Whangbo Pattern Recognition and Machine Learning Lab. Department of SoftwareGachon UniversitySeongnamGyeonggido13120Korea Riphah School of Computing&Innovation Faculty of ComputingRiphah International University Lahore CampusLahore54000Pakistan School of Computer Science National College of Business Administration and EconomicsLahore54000Pakistan Department of Computer Engineering Gachon UniversitySeongnam13557Korea

Cloud computing is providing IT services to its customer based on Service level agreements(SLAs).It is important for cloud service providers to provide reliable Quality of service(QoS)and to maintain SLAs *** service providers need to predict possible service violations before the emergence of an issue to perform remedial actions for *** users’major concerns;the factors for service reliability are based on response time,accessibility,availab.lity,and *** this paper,we,therefore,experiment with the parallel mutant-Particle swarm optimization(PSO)for the detection and predictions of QoS violations in terms of response time,speed,accessibility,and *** paper also compares Simple-PSO and Parallel *** simulation results,it is observed that the proposed Parallel MutantPSO solution for cloud QoS violation prediction achieves 94%accuracy which is many accurate results and is computationally the fastest technique in comparison of conventional PSO technique.

关键词： Accountability particle swarm optimization mutant particle swarm optimization quality of service service level agreement

来源：评论

学校读者我要写书评

暂无评论

Trainable Joint Bilateral Filters for Enhanced Prediction Stability in Low-dose CT

arXiv

引用

arXiv 2022年

作者： Wagner, Fabian Thies, Mareike Denzinger, Felix Gu, Mingxuan Patwari, Mayank Ploner, Stefan Maul, Noah Pfaff, Laura Huang, Yixing Maier, Andreas Pattern Recognition Lab. Friedrich-Alexander-Universität Erlangen-Nürnberg Erlangen91058 Germany

Low-dose computed tomography (CT) denoising algorithms aim to enable reduced patient dose in routine CT acquisitions while maintaining high image quality. Recently, deep learning (DL)-based methods were introduced, outperforming conventional denoising algorithms on this task due to their high model capacity. However, for the transition of DL-based denoising to clinical practice, these data-driven approaches must generalize robustly beyond the seen training data. We, therefore, propose a hybrid denoising approach consisting of a set of trainable joint bilateral filters (JBFs) combined with a convolutional DL-based denoising network to predict the guidance image. Our proposed denoising pipeline combines the high model capacity enabled by DL-based feature extraction with the reliability of the conventional JBF. The pipeline’s ability to generalize is demonstrated by training on abdomen CT scans without metal implants and testing on abdomen scans with metal implants as well as on head CT data. When embedding two well-established DL-based denoisers (RED-CNN/QAE) in our pipeline, the denoising performance is improved by 10%/82% (RMSE) and 3%/81% (PSNR) in regions containing metal and by 6%/78% (RMSE) and 2%/4% (PSNR) on head CT data, compared to the respective vanilla model. Concluding, the proposed trainable JBFs limit the error bound of deep neural networks to facilitate the applicability of DL-based denoisers in low-dose CT pipelines. © 2022, CC BY.

关键词： Computerized tomography

来源：评论

学校读者我要写书评

暂无评论

Lightweight Landslide Detection Method Based On Depth Separable Convolution And Double Self-Attention Mechanism *

Lightweight Landslide Detection Method Based On Depth Separa...

引用

IEEE International Symposium on Geoscience and Remote Sensing (IGARSS)

作者： Weibin Li Yuhui Kong Rongfang Wang Chunlei Huo Jiawei Chen Yi Niu School of Artificial Intelligence Xidian University Xi’an China Lab. of AI Hangzhou Institute of Technology of Xidian University Hangzhou China National Laboratory of Pattern Recognition Beijing China

The landslide detection methods using remote sensing images are mostly based on the traditional convolutional neural network model with high depth and complexity. The paper proposes a lightweight method based on Depth Separable Convolution and Double Self-Attention Mechanism (DSC-DSAM) for detecting landslides in remote sensing images. This method aims to reduce storage space and improve detection speed while maintaining accuracy. In our model, it starts with using a lightweight convolutional neural network model. Then, the dual self-attention mechanism is applied to improve the accuracy. The proposed method is compared with other existing classification models, and it is shown to have advantages in memory space and detection speed while maintaining accuracy.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Probabilistic Expression System for Fingerprint Identification Findings Based on Stability 7

A Probabilistic Expression System for Fingerprint Identifica...

引用

7th IEEE International Conference on Network Intelligence and Digital Content, IC-NIDC 2021

作者： Wang, Jingye Zheng, Yixiao Xiong, Weiyu Chen, Junhan Ma, Zhanyu Ma, Rongliang Beijing University of Posts and Telecommunications Pattern Recognition and Intelligent Systems Lab. Beijing100876 China Institute of Forensic Science Ministry of Public Security Beijing100045 China

ISBN: (纸本)9781665405829

Human fingerprints is of critical importance for law enforcement agencies to identify suspects. At present, the number of fingerprint matching feature points is generally used to determine whether two fingerprints are from the same person, and the criteria vary from country to country. This paper establishes a probabilistic expression system from a statistical point of view to give the probability of similarity between two fingerprints, which can be used to characterize the degree of similarity between the two fingerprints. Our method is now being tried for latent fingerprint and tenprints identification conclusion expression in public safety. © 2021 IEEE.

关键词： Palmprint recognition

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：