FuCiTNet: Improving the generalization of deep learning networks by the fusion of learned class-inherent transformations
Por favor, use este identificador para citar o enlazar este ítem:
http://hdl.handle.net/10045/107983
Título: | FuCiTNet: Improving the generalization of deep learning networks by the fusion of learned class-inherent transformations |
---|---|
Autor/es: | Rey-Area, Manuel | Guirado, Emilio | Tabik, Siham | Ruiz-Hidalgo, Javier |
Centro, Departamento o Servicio: | Universidad de Alicante. Instituto Multidisciplinar para el Estudio del Medio "Ramón Margalef" |
Palabras clave: | Deep neural networks | Generalization | Pre-processing | Transformation | GANs (Generative Adversarial Networks) | Classification | Small dataset |
Fecha de publicación: | nov-2020 |
Editor: | Elsevier |
Cita bibliográfica: | Information Fusion. 2020, 63: 188-195. doi:10.1016/j.inffus.2020.06.015 |
Resumen: | It is widely known that very small datasets produce overfitting in Deep Neural Networks (DNNs), i.e., the network becomes highly biased to the data it has been trained on. This issue is often alleviated using transfer learning, regularization techniques and/or data augmentation. This work presents a new approach, independent but complementary to the previous mentioned techniques, for improving the generalization of DNNs on very small datasets in which the involved classes share many visual features. The proposed model, called FuCiTNet (Fusion Class inherent Transformations Network), inspired by GANs, creates as many generators as classes in the problem. Each generator, k, learns the transformations that bring the input image into the k-class domain. We introduce a classification loss in the generators to drive the leaning of specific k-class transformations. Our experiments demonstrate that the proposed transformations improve the generalization of the classification model in three diverse datasets. |
Patrocinador/es: | This work partially supported by the Spanish Ministry of Science and Technology under the project TIN2017-89517-P and the project TEC2016-75976-R, financed by the Spanish Ministerio de Economía, Industria y Competitividad and the European Regional Development Fund (ERDF). S. Tabik was supported by the Ramon y Cajal Programme (RYC-2015-18136). E.G was supported by the European Research Council (ERC Grant agreement 647038 [BIODESERT]), with additional support from Generalitat Valenciana (CIDEGENT/2018/041). |
URI: | http://hdl.handle.net/10045/107983 |
ISSN: | 1566-2535 (Print) | 1872-6305 (Online) |
DOI: | 10.1016/j.inffus.2020.06.015 |
Idioma: | eng |
Tipo: | info:eu-repo/semantics/article |
Derechos: | © 2020 Elsevier B.V. |
Revisión científica: | si |
Versión del editor: | https://doi.org/10.1016/j.inffus.2020.06.015 |
Aparece en las colecciones: | Personal Investigador sin Adscripción a Grupo INV - DRYLAB - Artículos de Revistas |
Archivos en este ítem:
Archivo | Descripción | Tamaño | Formato | |
---|---|---|---|---|
Rey-Area_etal_2020_InformationFusion_final.pdf | Versión final (acceso restringido) | 2,7 MB | Adobe PDF | Abrir Solicitar una copia |
Rey-Area_etal_2020_InformationFusion_preprint.pdf | Preprint (acceso abierto) | 387,83 kB | Adobe PDF | Abrir Vista previa |
Todos los documentos en RUA están protegidos por derechos de autor. Algunos derechos reservados.