A study of the effect of noise and occlusion on the accuracy of convolutional neural networks applied to 3D object recognition
Por favor, use este identificador para citar o enlazar este ítem:
http://hdl.handle.net/10045/72633
Título: | A study of the effect of noise and occlusion on the accuracy of convolutional neural networks applied to 3D object recognition |
---|---|
Autor/es: | Garcia-Garcia, Alberto | Garcia-Rodriguez, Jose | Orts-Escolano, Sergio | Oprea, Sergiu | Gomez-Donoso, Francisco | Cazorla, Miguel |
Grupo/s de investigación o GITE: | Informática Industrial y Redes de Computadores | Robótica y Visión Tridimensional (RoViT) |
Centro, Departamento o Servicio: | Universidad de Alicante. Departamento de Tecnología Informática y Computación | Universidad de Alicante. Departamento de Ciencia de la Computación e Inteligencia Artificial | Universidad de Alicante. Instituto Universitario de Investigación Informática |
Palabras clave: | Deep learning | 3D object recognition | Convolutional neural networks | Noise | Occlusion | Caffe |
Área/s de conocimiento: | Arquitectura y Tecnología de Computadores | Ciencia de la Computación e Inteligencia Artificial |
Fecha de publicación: | nov-2017 |
Editor: | Elsevier |
Cita bibliográfica: | Computer Vision and Image Understanding. 2017, 164: 124-134. doi:10.1016/j.cviu.2017.06.006 |
Resumen: | In this work, we carry out a study of the effect of adverse conditions, which characterize real-world scenes, on the accuracy of a Convolutional Neural Network applied to 3D object class recognition. Firstly, we discuss possible ways of representing 3D data to feed the network. In addition, we propose a set of representations to be tested. Those representations consist of a grid-like structure (fixed and adaptive) and a measure for the occupancy of each cell of the grid (binary and normalized point density). After that, we propose and implement a Convolutional Neural Network for 3D object recognition using Caffe. At last, we carry out an in-depth study of the performance of the network over a 3D CAD model dataset, the Princeton ModelNet project, synthetically simulating occlusions and noise models featured by common RGB-D sensors. The results show that the volumetric representations for 3D data play a key role on the recognition process and Convolutional Neural Network can be considerably robust to noise and occlusions if a proper representation is chosen. |
Patrocinador/es: | This work has been supported by the Spanish Government DPI2013-40534-R grant for the SIRMAVED project, also supported with FEDER funds. This work has also been funded by the grant “Ayudas para Estudios de Máster e Iniciación a la Investigación” from the University of Alicante. |
URI: | http://hdl.handle.net/10045/72633 |
ISSN: | 1077-3142 (Print) | 1090-235X (Online) |
DOI: | 10.1016/j.cviu.2017.06.006 |
Idioma: | eng |
Tipo: | info:eu-repo/semantics/article |
Derechos: | © 2017 Elsevier Inc. |
Revisión científica: | si |
Versión del editor: | http://dx.doi.org/10.1016/j.cviu.2017.06.006 |
Aparece en las colecciones: | INV - I2RC - Artículos de Revistas INV - RoViT - Artículos de Revistas INV - AIA - Artículos de Revistas |
Archivos en este ítem:
Archivo | Descripción | Tamaño | Formato | |
---|---|---|---|---|
2017_Garcia-Garcia_etal_CompVisImageUnderst_final.pdf | Versión final (acceso restringido) | 2,74 MB | Adobe PDF | Abrir Solicitar una copia |
2017_Garcia-Garcia_etal_CompVisImageUnderst_preprint.pdf | Preprint (acceso abierto) | 3,62 MB | Adobe PDF | Abrir Vista previa |
Todos los documentos en RUA están protegidos por derechos de autor. Algunos derechos reservados.