检索结果-内蒙古大学图书馆

2022 Findings of the Association for Computational Linguistics: EMNLP 2022

作者： Li, Lei Lin, Yankai Ren, Xuancheng Zhao, Guangxiang Li, Peng Zhou, Jie Sun, Xu MOE Key Lab of Computational Linguistics School of Computer Science Peking University China Gaoling School of Artificial Intelligence Renmin University of China Beijing China Beijing Key Laboratory of Big Data Management and Analysis Methods Beijing China Tsinghua University China Pattern Recognition Center WeChat AI Tencent Inc. China

Investigating better ways to reuse the released pre-trained language models (PLMs) can significantly reduce the computational cost and the potential environmental side-effects. This paper explores a novel PLM reuse paradigm, Knowledge Integration (KI). Without human annotations available, KI aims to merge the knowledge from different teacher-PLMs, each of which specializes in a different classification problem, into a versatile student model. To achieve this, we first derive the correlation between virtual golden supervision and teacher predictions. We then design a Model Uncertainty-aware Knowledge Integration (MUKI) framework to recover the golden supervision for the student. Specifically, MUKI adopts Monte-Carlo Dropout to estimate model uncertainty for the supervision integration. An instance-wise re-weighting mechanism based on the margin of uncertainty scores is further incorporated, to deal with the potential conflicting supervision from teachers. Experimental results demonstrate that MUKI achieves substantial improvements over baselines on benchmark datasets. Further analysis shows that MUKI can generalize well for merging teacher models with heterogeneous architectures, and even teachers major in cross-lingual datasets. © 2022 Association for Computational Linguistics.

关键词： Integration

来源：评论

学校读者我要写书评

暂无评论

Comparative Analysis of Radiomic Features and Gene Expression Profiles in Histopathology Data Using Graph Neural Networks

arXiv

引用

arXiv 2023年

作者： Monroy, Luis Carlos Rivera Rist, Leonhard Eberhardt, Martin Ostalecki, Christian Bauer, Andreas Vera, Julio Breininger, Katharina Maier, Andreas Pattern Recognition Lab Friedrich-Alexander-Universität Erlangen-Nürnberg Erlangen Germany Department of Dermatology Universitätsklinikum Erlangen Erlangen Germany Department Artificial Intelligence in Biomedical Engineering FAU Erlangen-Nürnberg Erlangen Germany

This study leverages graph neural networks to integrate MELC data with Radiomic-extracted features for melanoma classification, focusing on cellwise analysis. It assesses the effectiveness of gene expression profiles and Radiomic features, revealing that Radiomic features, particularly when combined with UMAP for dimensionality reduction, significantly enhance classification performance. Notably, using Radiomics contributes to increased diagnostic accuracy and computational efficiency, as it allows for the extraction of critical data from fewer stains, thereby reducing operational costs. This methodology marks an advancement in computational dermatology for melanoma cell classification, setting the stage for future research and potential developments. © 2023, CC BY.

关键词： Dermatology

来源：评论

学校读者我要写书评

暂无评论

Employing Graph Representations for Cell-Level Characterization of Melanoma MELC Samples

Employing Graph Representations for Cell-Level Characterizat...

引用

IEEE International Symposium on Biomedical Imaging

作者： Luis Carlos Rivera Monroy Leonhard Rist Martin Eberhardt Christian Ostalecki Andreas Baur Julio Vera Katharina Breininger Andreas Maier Pattern Recognition Lab Friedrich-Alexander-Universität Erlangen-Nürnberg Erlangen Germany Department of Dermatology Universitätsklinikum Erlangen Erlangen Germany Department Artificial Intelligence in Biomedical Engineering FAU Erlangen-Nürnberg Erlangen Germany

Histopathology imaging is crucial for the diagnosis and treatment of skin diseases. For this reason, computer-assisted approaches have gained popularity and shown promising results in tasks such as segmentation and classification of skin disorders. However, collecting essential data and sufficiently high-quality annotations is a challenge. This work describes a pipeline that uses suspected melanoma samples that have been characterized using Multi-Epitope-Ligand Cartography (MELC). This cellular level characterization of the tissue is then represented as a graph and finally used to train a graph neural network. This imaging technology, combined with the methodology proposed in this work, achieves a classification accuracy of 87 %, outperforming existing approaches by 10 %.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Rice Leaves Disease Diagnose Empowered with Transfer Learning

引用

Computer Systems Science & Engineering 2022年第9期42卷 1001-1014页

作者： Nouh Sabri Elmitwally Maria Tariq Muhammad Adnan Khan Munir Ahmad Sagheer Abbas Fahad Mazaed Alotaibi College of Computer and Information Sciences Jouf UniversitySakaka72341Saudi Arabia Department of Computer Science Faculty of Computers and Artificial IntelligenceCairo University12613Egypt School of Computer Science National College of Business Administration and EconomicsLahore54000Pakistan Department of Computer Science Lahore Garrison UniversityLahore54000Pakistan Pattern Recognition and Machine Learning Lab Department of SoftwareGachon UniversitySeongnam13557Korea Department of Information System Faculty of Computing and Information TechnologyKing Abdulaziz UniversityJeddah21589Saudi Arabia

In the agricultural industry,rice infections have resulted in significant productivity and economic *** infections must be recognized early on to regulate and mitigate the effects of the *** diagnosis of disease severity effects or incidence can preserve production from quantitative and qualitative losses,reduce pesticide use,and boost ta country’s *** the health of a rice plant through its leaves is usually done as a manual ocular *** this manuscript,three rice plant diseases:Bacterial leaf blight,Brown spot,and Leaf smut,were identified using the Alexnet *** research shows that any reduction in rice plants will have a significant beneficial impact on alleviating global food hunger by increasing supply,lowering prices,and reducing production's environmental impact that affects the economy of any *** would be able to get more exact and faster results with this technology,allowing them to administer the most acceptable treatment *** Using Alex Net,the proposed approach achieved a 99.0%accuracy rate for diagnosing rice leaves disease.

关键词： Rice bacterial leaf blight brown spot leaf smut machine learning alexnet

来源：评论

学校读者我要写书评

暂无评论

GeoPix: Multi-Modal Large Language Model for Pixel-level Image Understanding in Remote Sensing

arXiv

引用

arXiv 2025年

作者： Ou, Ruizhe Hu, Yuan Zhang, Fan Chen, Jiaxin Liu, Yu Pattern Recognition and Intelligent System Lab School of Artificial Intelligence Beijing University of Posts and Telecommunications Beijing100876 China Institute of Remote Sensing and Geographic Information Systems School of Earth and Space Sciences Peking University Beijing100871 China Peking University Ordos Research Institute of Energy Ordos017000 China

Multi-modal large language models (MLLMs) have achieved remarkable success in image- and region-level remote sensing (RS) image understanding tasks, such as image captioning, visual question answering, and visual grounding. However, existing RS MLLMs lack the pixel-level dialogue capability, which involves responding to user instructions with segmentation masks for specific instances. In this paper, we propose GeoPix, a RS MLLM that extends image understanding capabilities to the pixel level. This is achieved by equipping the MLLM with a mask predictor, which transforms visual features from the vision encoder into masks conditioned on the LLM’s segmentation token embeddings. To facilitate the segmentation of multi-scale objects in RS imagery, a class-wise learnable memory module is integrated into the mask predictor to capture and store class-wise geo-context at the instance level across the entire dataset. In addition, to address the absence of large-scale datasets for training pixel-level RS MLLMs, we construct the GeoPixInstruct dataset, comprising 65,463 images and 140,412 instances, with each instance annotated with text descriptions, bounding boxes, and masks. Furthermore, we develop a two-stage training strategy to balance the distinct requirements of text generation and masks prediction in multi-modal multi-task optimization. Extensive experiments verify the effectiveness and superiority of GeoPix in pixel-level segmentation tasks, while also maintaining competitive performance in image- and region-level benchmarks. The models, dataset, and code are publicly available at https://***/Norman-Ou/GeoPix. © 2025, CC BY.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Enhancing Cross-Modality Synthesis: Subvolume Merging for MRI-to-CT Conversion

arXiv

引用

arXiv 2024年

作者： Fan, Fuxin Qiu, Jingna Huang, Yixing Maier, Andreas The Pattern Recognition Lab Friedrich-Alexander-Universität Erlangen Nürnberg Germany The Department of Artificial Intelligence in Biomedical Engineering Friedrich-Alexander-Universität Erlangen Nürnberg Germany The Department of Radiation Oncology University Hospital Erlangen Friedrich-Alexander-Universität Erlangen Nürnberg Germany

Providing more precise tissue attenuation information, synthetic computed tomography (sCT) generated from magnetic resonance imaging (MRI) contributes to improved radiation therapy treatment planning. In our study, we employ the advanced SwinUNETR framework for synthesizing CT from MRI images. Additionally, we introduce a three-dimensional subvolume merging technique in the prediction process. By selecting an optimal overlap percentage for adjacent subvolumes, stitching artifacts are effectively mitigated, leading to a decrease in the mean absolute error (MAE) between sCT and the labels from 52.65 HU to 47.75 HU. Furthermore, implementing a weight function with a gamma value of 0.9 results in the lowest MAE within the same overlap area. By setting the overlap percentage between 50% and 70%, we achieve a balance between image quality and computational efficiency. © 2024, CC BY.

关键词： Computerized tomography

来源：评论

学校读者我要写书评

暂无评论

Personality Detection Using Context Based Emotions in Cognitive Agents

引用

Computers, Materials & Continua 2022年第3期70卷 4947-4964页

作者： Nouh Sabri Elmitwally Asma Kanwal Sagheer Abbas Muhammad A.Khan Muhammad Adnan Khan Munir Ahmad Saad Alanazi College of Computer and Information Sciences Jouf UniversitySakaka72341Saudi Arabia Department of Computer Science Faculty of Computers and Artificial IntelligenceCairo University12613Egyp t 3School of Computer Science National College of Business administration and EconomicsLahore54000Pakistan Department of Computer Science GC UniversityLahorePakistan Riphah School of Computing&Innovation Faculty of ComputingRiphah International UniversityLahore CampusLahore54000Pakistan Pattern Recognition and Machine Learning Lab Department of SoftwareGachon UniversitySeongnam13120Korea

Detection of personality using emotions is a research domain in artificial *** present,some agents can keep the human’s profile for interaction and adapts themselves according to their ***,the effective method for interaction is to detect the person’s personality by understanding the emotions and context of the *** idea behind adding personality in cognitive agents begins an attempt to maximize adaptability on the basis of *** our daily life,humans socially interact with each other by analyzing the emotions and context of interaction from audio or visual *** paper presents a conceptual personality model in cognitive agents that can determine personality and behavior based on some text input,using the context subjectivity of the given data and emotions obtained from a particular situation/*** proposed work consists of Jumbo Chatbot,which can chat with *** this social interaction,the chatbot predicts human personality by understanding the emotions and context of interactive ***,the Jumbo chatbot is using the BFI technique to interact with a *** accuracy of proposed work varies and improve through getting more experiences of interaction.

关键词： Emotions fuzzy personality detection contextual analysis semantic analysis

来源：评论

学校读者我要写书评

暂无评论

Improved HER2 Tumor Segmentation with Subtype Balancing Using Deep Generative Networks

Improved HER2 Tumor Segmentation with Subtype Balancing Usin...

引用

IEEE International Symposium on Biomedical Imaging

作者： Mathias Öttl Jana Steenpass Matthias Rübner Carol I. Geppert Jingna Qiu Frauke Wilm Arndt Hartmann Matthias W. Beckmann Peter A. Fasching Andreas Maier Ramona Erber Katharina Breininger Pattern Recognition Lab Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) Germany Institute of Pathology University Hospital Erlangen FAU Germany Department of Gynecology and Obstetrics University Hospital Erlangen FAU Germany Department Artificial Intelligence in Biomedical Engineering FAU Germany

Tumor segmentation in histopathology images is often complicated by its composition of different histological subtypes and class imbalance. Oversampling subtypes with low prevalence features is not a satisfactory solution since it eventually leads to overfitting. We propose to create synthetic images with semantically-conditioned deep generative networks and to combine subtype-balanced synthetic images with the original dataset to achieve better segmentation performance. We show the suitability of Generative Adversarial Networks (GANs) and especially diffusion models to create realistic images based on subtype-conditioning for the use case of HER2-stained histopathology. Additionally, we show the capability of diffusion models to conditionally inpaint HER2 tumor areas with modified subtypes. Combining the original dataset with the same amount of diffusion-generated images increased the tumor Dice score from 0.833 to 0.854 and almost halved the variance between the HER2 subtype recalls. These results create the basis for more reliable automatic HER2 analysis with lower performance variance between individual HER2 subtypes.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Handling label Uncertainty on the Example of Automatic Detection of Shepherd’s Crook RCA in Coronary CT Angiography

Handling Label Uncertainty on the Example of Automatic Detec...

引用

IEEE International Symposium on Biomedical Imaging

作者： Felix Denzinger Michael Wels Oliver Taubmann Florian Kordon Fabian Wagner Stephanie Mehltretter Mehmet A. Gülsün Max Schöbinger Florian André Sebastian Buß Johannes Görich Michael Sühling Andreas Maier Katharina Breininger Pattern Recognition Lab FAU Erlangen-Nürnberg Erlangen Germany Computed Tomography Siemens Healthcare GmbH Forchheim Germany Das Radiologische Zentrum Sinsheim-Eberbach-Erbach-Walldorf-Heidelberg Germany Department Artificial Intelligence in Biomedical Engineering FAU Erlangen-Nürnberg

Coronary artery disease (CAD) is often treated minimally invasively with a catheter being inserted into the diseased coronary vessel. If a patient exhibits a Shepherd’s Crook (SC) Right Coronary Artery (RCA) – an anatomical norm variant of the coronary vasculature – the complexity of this procedure is increased. Automated reporting of this variant from coronary CT angiography screening would ease prior risk assessment. We propose a 1D convolutional neural network which leverages a sequence of residual dilated convolutions to automatically determine this norm variant from a prior extracted vessel centerline. As the SC RCA is not clearly defined with respect to concrete measurements, labeling also includes qualitative aspects. Therefore, 4.23 % samples in our dataset of 519 RCA centerlines were labeled as unsure SC RCAs, with 5.97 % being labeled as sure SC RCAs. We explore measures to handle this label uncertainty, namely global/model-wise random assignment, exclusion, and soft label assignment. Furthermore, we evaluate how this uncertainty can be leveraged for the determination of a rejection class. With our best configuration, we reach an area under the receiver operating characteristic curve (AUC) of 0.938 on confident labels. Moreover, we observe an increase of up to 0.020 AUC when rejecting 10 % of the data and leveraging the labeling uncertainty information in the exclusion process.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Ra-Mmir: Multi-Modal Image Registration by Robust Adaptive Variation Attention Gauge Field

SSRN

引用

SSRN 2023年

作者： Qiu, Junhui Li, Hao Cao, Hualong Zhai, Xiangshuai Liu, Xuedong Sang, Meng Yu, Kailong Sun, Yunpin Yang, Yang Tan, Pan Yunnan Normal University School of Information Kunming650504 China Yunnan Normal University Pattern Recognition and Artificial Intelligence Lab Kunming650504 China Wuhan University School of Computer Science Wuhan430072 China Renmin University of China School of Information Beijing100872 China

Multi-modal image registration is widely used in high-level vision tasks. In adverse conditions, high-level visual tasks often emphasize the analysis of important target regions in the image, however precise registration of multi-modal images remains a challenge in adverse conditions. To address this issue, we rethought the collaboration between image registration and high-level visual tasks and proposed a robust adaptive attention gauge field registration strategy that allows for flexible attention to both the target regions and global areas features present in the image. To focus on key feature extraction and registration of target regions and global areas flexibly in adverse conditions, we have designed a feature fusion map with weak boundaries and probabilistic properties, which unifies feature extraction and description, while guiding the network to flexibly extract and match features in both target regions and global context, this is referred to as Attention-Gauge Field Encoding. To improve the robustness and timeliness of the algorithm in extracting and describing target regions or global areas features in adverse conditions, we proposed a robust adaptive probability distribution and constructed a robust adaptive attention for building the gauge field. To achieve global optimal robust adaptive parameters, we designed a simulated annealing-like method during training. To make the spatial transformation model better fit feature matching and image registration, we designed a deep learning and modeling approach based on spatial similarity metrics. In experiments on high-level visual tasks such as image fusion, object detection, and 3D reconstruction, our method had the best collaborative performance and showed the best registration results under adverse conditions. © 2023, The Authors. All rights reserved.

关键词： Feature extraction

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：