RT info:eu-repo/semantics/article
T1 Cross Validation Voting for Improving CNN Classification in Grocery Products
A1 Duque Domingo, Jaime
A1 Medina Aparicio, Roberto
A1 González Rodrigo, Luis Miguel
K1 Artificial intelligence
K1 Cross-validation-voting, CVV, voting, boosting, ResNeXt, EfficientNet, Wide ResNet, CNN, grocery image classification.
K1 Artificial intelligence
AB The development of deep neural networks that has been carried out in recent years allowssolving highly complex computer vision classification problems. Often, although the results obtained withthese classifiers are high, there are certain sectors that seek greater accuracy from these systems. Increasing the accuracy of neural networks can be achieved through ensemble learning, which combines different classifiers with the aim of selecting a winner based on different criteria about them. These techniques have traditionally shown good results although they involve training models of different nature and can even produce an overfitting with respect to the training data, so datasets must be chosen to correctly evaluate the result. In this paper, a Cross-Validation-Voting (CVV) technique for grocery product classification is presented. This technique improves several single state-of-the-art classifiers without combining different ones and avoids the problems of overfitting with respect to the training set. The single classifiers are trained multiple times against distributed sets to show how the results obtained to date from the classification of a well-known dataset are improved. In this dataset, an extensive test set was previously selected by the authors to show comparable results with other papers in the literature. The technique is valid not only for vision nets and can be used to solve numerous problems with different kinds of neural networks and classifiers.
PB IEEE
SN 2169-3536
YR 2022
FD 2022-02-08
LK https://uvadoc.uva.es/handle/10324/65739
UL https://uvadoc.uva.es/handle/10324/65739
LA spa
NO IEEE Access, Febrero 2022, vol. 10, p. 20913-20925
NO Producción Científica
DS UVaDOC
RD 24-nov-2024