Adaptive convolution kernel for artificial neural networks
Yükleniyor...
Dosyalar
Tarih
2021-02
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Academic Press Inc.
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
Many deep neural networks are built by using stacked convolutional layers of fixed and single size (often 3 × 3) kernels. This paper describes a method for learning the size of convolutional kernels to provide varying size kernels in a single layer. The method utilizes a differentiable, and therefore backpropagation-trainable Gaussian envelope which can grow or shrink in a base grid. Our experiments compared the proposed adaptive layers to ordinary convolution layers in a simple two-layer network, a deeper residual network, and a U-Net architecture. The results in the popular image classification datasets such as MNIST, MNIST-CLUTTERED, CIFAR-10, Fashion, and ‘‘Faces in the Wild’’ showed that the adaptive kernels can provide statistically significant improvements on ordinary convolution kernels. A segmentation experiment in the Oxford-Pets dataset demonstrated that replacing ordinary convolution layers in a U-shaped network with 7 × 7 adaptive layers can improve its learning performance and ability to generalize.
Açıklama
This work was supported by The Scientific and Technological Research Council of Turkey programme (TUBITAK-1001 no: 118E722), Isik University BAP programme, Turkey (no: 16A202), and NVIDIA hardware donation of a Tesla K40 GPU unit, Turkey.
Anahtar Kelimeler
Adaptive convolution, Image classification, Multi-scale convolution, Residual networks, Backpropagation, Classification (of information), Convolution, Deep learning, Deep neural networks, Image enhancement, Learning systems, Network layers, Adaptive kernels, Classification datasets, Convolution kernel, Convolutional kernel, Gaussian envelope, Learning performance, NET architecture, Two-layer network, Multilayer neural networks
Kaynak
Journal of Visual Communication and Image Representation
WoS Q Değeri
Q2
Scopus Q Değeri
Q1
Cilt
75
Sayı
Künye
Tek, F. B., Çam, İ. & Karlı, D. (2021). Adaptive convolution kernel for artificial neural networks. Journal of Visual Communication and Image Representation, 75, 1-11.doi:10.1016/j.jvcir.2020.103015
