Please use this identifier to cite or link to this item: http://hdl.handle.net/10553/58308
Title: Automatic segmentation based on deep learning techniques for diabetic foot monitoring through multimodal images
Authors: Hernández Guedes, Abián 
Arteaga-Marrero, Natalia
Villa, Enrique
Fabelo Gómez, Himar Antonio 
Marrero Callicó, Gustavo Iván 
Ruiz Alzola, Juan Bautista 
UNESCO Clasification: 3314 Tecnología médica
Keywords: RGB-D images
Multimodal images
Deep Learning
Automatic segmentation
Issue Date: 2019
Publisher: Springer 
Journal: Lecture Notes in Computer Science 
Conference: 20th International Conference on Image Analysis and Processing, (ICIAP 2019) 
Abstract: Temperature data acquired by infrared sensors provide relevant information to assess different medical pathologies in early stages, when the symptoms of the diseases are not visible yet to the naked eye. Currently, a clinical system that exploits the use of multimodal images (visible, depth and thermal infrared) is being developed for diabetic foot monitoring. The workflow required to analyze these images starts with their acquisition and the automatic feet segmentation. A novel approach is presented for automatic feet segmentation using Deep Learning employing an architecture composed of an encoder and decoder (U-Net architecture) and applying a segmentation of planes in point cloud data, using the depth information of pixels labeled in the neural network prediction. The proposed automatic segmentation is a robust method for this case study, providing results in a short time and achieving better performance than other traditional segmentation methods as well as a basic U-Net segmentation system.
URI: http://hdl.handle.net/10553/58308
ISBN: 978-3-030-30644-1
ISSN: 0302-9743
DOI: 10.1007/978-3-030-30645-8_38
Source: Image Analysis and Processing – ICIAP 2019. ICIAP 2019. Lecture Notes in Computer Science, v. 11752 LNCS, p. 414-424
Appears in Collections:Capítulo de libro
Show full item record

SCOPUSTM   
Citations

7
checked on Nov 3, 2024

WEB OF SCIENCETM
Citations

6
checked on Nov 3, 2024

Page view(s)

214
checked on Nov 1, 2024

Google ScholarTM

Check

Altmetric


Share



Export metadata



Items in accedaCRIS are protected by copyright, with all rights reserved, unless otherwise indicated.