Supervised and unsupervised learning approaches for the labeling of multivariate images

机译：有监督和无监督学习方法，用于标注多元图像

获取原文

获取原文并翻译 | 示例

页面导航

摘要
著录项
相似文献
相关主题

摘要

Abstract: A multivariate numeric image can be seen as a 3-waydata table: two dimensions of this table are of spatialnature whereas the other characterizes the constitutiveunivariate images. The process of labeling consists inassigning a qualitative group to each pixel of theoriginal multivariate image. A supervised learningmethod, stepwise discriminant analysis was comparedwith two unsupervised methods, simple C-meansclustering (CMC) and fuzzy C-means. As illustrativeexample, the methods were applied on multivariateimages of sections of maize kernels obtained byfluorescence imaging. CMC requires the utilization of afunction assessing the distance between somerepresentative patterns and the pixel vectors. Therelative interest of Euclidean distance and Mahalanobisdistance was investigated. The best results wereobtained by using CMC and simple Euclidean distance. Inthese conditions, it was possible to identify, with noa priori knowledge, the main tissues of maize. !6

机译：摘要：多元数值图像可以看作是3维数据表：该表的两个维度具有空间性质，而另一个则表示本构单变量图像。标记过程包括为原始多元图像的每个像素分配一个定性组。将监督学习方法，逐步判别分析与两种无监督方法进行了比较，即简单C均值聚类（CMC）和模糊C均值。作为说明性实例，该方法被应用于通过荧光成像获得的玉米粒的切片的多变量图像。 CMC需要利用功能来评估某些代表性图案与像素向量之间的距离。研究了欧氏距离和马氏距离的相对兴趣。通过使用CMC和简单的欧几里德距离获得了最佳结果。在这种情况下，无需先验就可以鉴定玉米的主要组织。！6

著录项

来源
《Precision Agriculture and Biological Quality》|1998年|p.44-52|共9页
会议地点 Boston MA(US)
作者
Author(s): Dominique Bertrand Institut National de la Recherche Agronomique Nantes Cedex 3 France; Bruno Novales Institut National de la Recherche Agronomique Nantes Cedex 3 France; Younes Chtioui Institut National de la Recherche Agronomique Farg;
展开▼
作者单位

展开▼
会议组织
原文格式 PDF
正文语种 eng
中图分类 S-;
关键词

相似文献

外文文献
中文文献
专利

1. Deep-learning-based image quality enhancement of compressed sensing magnetic resonance imaging of vessel wall: comparison of self-supervised and unsupervised approaches [J] . Da-in Eun, Ryoungwoo Jang, Woo Seok Ha, Scientific reports. . 2020,第1期

机译：基于深度学习的图像质量增强血管壁压缩传感磁共振成像：自我监督和无监督方法比较
2. MaskMitosis: a deep learning framework for fully supervised, weakly supervised, and unsupervised mitosis detection in histopathology images [J] . Medical and Biological Engineering and Computing: Journal of the International Federation for Medical and Biological Engineering . 2020,第7期

机译：面具症：在组织病理学图像中完全监督，弱弱监督和无调节丝分裂检测的深度学习框架
3. Lung and Pancreatic Tumor Characterization in the Deep Learning Era: Novel Supervised and Unsupervised Learning Approaches [J] . Hussein Sarfaraz, Kandel Pujan, Bolan Candice W., IEEE Transactions on Medical Imaging . 2019,第8期

机译：深度学习时代的肺和胰腺肿瘤表征：新型的有监督和无监督学习方法
4. Supervised and unsupervised learning approaches for the labeling ofmultivariate images, [C] . Dominique Bertrand, Institut National de la Recherche Agronomique, Nantes Cedex, Precision Agriculture and Biological Quality . 1998

机译：有监督和无监督的标注多元图像的学习方法，
5. Learning from partially labeled data: Unsupervised and semi-supervised learning on graphs and learning with distribution shifting. [D] . Huang, Jiayuan. 2007

机译：从部分标记的数据中学习：在图上进行无监督和半监督学习，并通过分布转移进行学习。
6. Comparison of Supervised and Unsupervised Deep Learning Methods for Medical Image Synthesis between Computed Tomography and Magnetic Resonance Images [O] . Yafen Li, Wen Li, Jing Xiong, 2020

机译：计算机断层扫描与磁共振图像中医学图像综合的监督和无监督深度学习方法的比较
7. Deep-learning-based image quality enhancement of compressed sensing magnetic resonance imaging of vessel wall: comparison of self-supervised and unsupervised approaches [O] . Da-in Eun, Ryoungwoo Jang, Woo Seok Ha, 2020

机译：基于深度学习的图像质量增强血管壁压缩传感磁共振成像：自我监督和无监督方法比较

Supervised and unsupervised learning approaches for the labeling of multivariate images

摘要

著录项

相似文献

相关主题

期刊订阅