Analysis of single‑ and dual‑dictionary strategies in pedestrian classification
![Thumbnail](/xmlui/bitstream/handle/10234/175078/58818.pdf.jpg?sequence=6&isAllowed=y)
View/ Open
Metadata
Show full item recordcomunitat-uji-handle:10234/9
comunitat-uji-handle2:10234/43662
comunitat-uji-handle3:10234/43643
comunitat-uji-handle4:
INVESTIGACIONMetadata
Title
Analysis of single‑ and dual‑dictionary strategies in pedestrian classificationDate
2018Publisher
Springer VerlagISSN
1433-7541; 1433-755XBibliographic citation
Traver, V.J. & Serra-Toro, C. Pattern Anal Applic (2018). https://doi.org/10.1007/s10044-018-0704-5Type
info:eu-repo/semantics/articlePublisher version
https://link.springer.com/article/10.1007/s10044-018-0704-5Version
info:eu-repo/semantics/acceptedVersionSubject
Abstract
Sparse coding has recently been a hot topic in visual tasks in image processing and computer vision. It has applications
and brings benefts in reconstruction-like tasks and in classifcation-like tasks as well. However, ... [+]
Sparse coding has recently been a hot topic in visual tasks in image processing and computer vision. It has applications
and brings benefts in reconstruction-like tasks and in classifcation-like tasks as well. However, regarding binary classifcation
problems, there are several choices to learn and use dictionaries that have not been studied. In particular, how
single-dictionary and dual-dictionary approaches compare in terms of classifcation performance is largely unexplored. We
compare three single-dictionary strategies and two dual-dictionary strategies for the problem of pedestrian classifcation
(“pedestrian” vs “background” images). In each of these fve cases, images are represented as the sparse coefcients induced
from the respective dictionaries, and these coefcients are the input to a regular classifer both for training and subsequent
classifcation of novel unseen instances. Experimental results with the INRIA pedestrian dataset suggest, on the one hand,
that dictionaries learned from only one of the classes, even from the background class, are enough for obtaining competitive
good classifcation performance. On the other hand, while better performance is generally obtained when instances of both
classes are used for dictionary learning, the representation induced by a single dictionary learned from a set of instances
from both classes provides comparable or even superior performance over the representations induced by two dictionaries
learned separately from the pedestrian and background classes. [-]
Is part of
Pattern Anal Applic (2018)Investigation project
TIN2013-46522-P ; PROMETEOII/2014/062Rights
© Springer-Verlag London Ltd., part of Springer Nature 2018.
“This is a post-peer-review, pre-copyedit version of an article published in Pattern Analysis and Applications. The final authenticated version is available online at: https://doi.org/10.1007/s10044-018-0704-5".
http://rightsstatements.org/vocab/InC/1.0/
info:eu-repo/semantics/openAccess
http://rightsstatements.org/vocab/InC/1.0/
info:eu-repo/semantics/openAccess
This item appears in the folowing collection(s)
- INIT_Articles [754]