@InProceedings{NazareCostMellPont:2018:EmAnUs,
author = "Nazare, Tiago S. and Costa, Gabriel B. Paranhos da and Mello,
Rodrigo F. de and Ponti, Moacir A.",
affiliation = "{University of S{\~a}o Paulo} and {University of S{\~a}o Paulo}
and {University of S{\~a}o Paulo} and {University of S{\~a}o
Paulo}",
title = "Color quantization in transfer learning and noisy scenarios: an
empirical analysis using convolutional networks",
booktitle = "Proceedings...",
year = "2018",
editor = "Ross, Arun and Gastal, Eduardo S. L. and Jorge, Joaquim A. and
Queiroz, Ricardo L. de and Minetto, Rodrigo and Sarkar, Sudeep and
Papa, Jo{\~a}o Paulo and Oliveira, Manuel M. and Arbel{\'a}ez,
Pablo and Mery, Domingo and Oliveira, Maria Cristina Ferreira de
and Spina, Thiago Vallin and Mendes, Caroline Mazetto and Costa,
Henrique S{\'e}rgio Gutierrez and Mejail, Marta Estela and Geus,
Klaus de and Scheer, Sergio",
organization = "Conference on Graphics, Patterns and Images, 31. (SIBGRAPI)",
publisher = "IEEE Computer Society",
address = "Los Alamitos",
keywords = "Deep learning, transfer learning, convolutional neural networks,
computer vision.",
abstract = "Transfer learning is seen as one of the most promising areas of
machine learning. Lately, features from pre-trained models have
been used to achieve state-of-the-art results in several machine
vision problems. Those models are usually employed when the
problem of interest does not have enough supervised examples to
support the network training from scratch. Most applications use
networks pre-trained on noise-free RGB image datasets, what is
observed even when the target domain counts on grayscale images or
when data is degraded by noise. In this paper, we evaluate the use
of Convolutional Neural Networks (CNNs) on such transfer learning
scenarios and the impact of using RGB trained networks on
grayscale image tasks. Our results confirm that the use of
networks trained using colored images on grayscale tasks hinders
the overall performance when compared to a similar network trained
on a quantized version of the original dataset. Results also show
that higher quantization levels (resulting in less colors)
increase the robustness of CNN features in the presence of
noise.",
conference-location = "Foz do Igua{\c{c}}u, PR, Brazil",
conference-year = "29 Oct.-1 Nov. 2018",
doi = "10.1109/SIBGRAPI.2018.00055",
url = "http://dx.doi.org/10.1109/SIBGRAPI.2018.00055",
language = "en",
ibi = "8JMKD3MGPAW/3RRA45S",
url = "http://urlib.net/ibi/8JMKD3MGPAW/3RRA45S",
targetfile = "SIB_2018.pdf",
urlaccessdate = "2024, May 02"
}