RT Journal Article
T1 U-Net/ResNet-50 Network with Transfer Learning for Semantic Segmentation in Search and Rescue.
A1 Salas Espinales, Andrés
A1 Vélez-Chávez, Elián
A1 Vázquez-Martín, Ricardo
A1 García-Cerezo, Alfonso José
A1 Mandow, Anthony
K1 Aprendizaje automático (Inteligencia artificial)
K1 Rescate
AB Semantic segmentation has been successfully adopted for scenarios such as indoor, outdoor, urban scenes, and synthetic scenes, but applications with scarce labeled data such as search-and-rescue (SAR), have not been addressed. In this work, we propose a transfer learning approach where the U-Net convolutional neural network incorporates ResNet-50 as an encoder for the segmentation of objects in SAR situations. First, the proposed model is trained and validated with 19 classes of the CityScapes dataset. Then we test the proposed approach by i) training the model with a set of 14 Cityscapes classes with relevant similarities to SAR classes, and ii) using transfer learning with the self-developed dataset in SAR scenarios, which has 349 semantic labeled SAR images. The results indicate good recognition in classes with significant presence on the training images.
PB Springer Nature
YR 2024
FD 2024-04-27
LK https://hdl.handle.net/10630/36759
UL https://hdl.handle.net/10630/36759
LA eng
NO Salas-Espinales A., Vélez-Chávez E., Vázquez-Martín R., García-Cerezo A., Mandow A. U-Net/ResNet-50 Network with Transfer Learning for Semantic Segmentation in Search and Rescue (2024) Lecture Notes in Networks and Systems, 978 LNNS, pp. 244 - 255 DOI: https://www.doi.org/10.1007/978-3-031-59167-9_21
NO https://www.springernature.com/gp/open-science/policies/book-policies
NO This work has been partially funded by the Ministerio de Ciencia, Innovación y Universidades, Gobierno de España, project {\small PID2021-122944OB-I00}. The first author received a grant from Asociación Universitaria Iberoamericana de Postgrado (AUIP), Universidad de Málaga, and Universidad Técnica de Manabí.
DS RIUMA. Repositorio Institucional de la Universidad de Málaga
RD 22 ene 2026