Skip to main content
Fig. 1 | BMC Neuroscience

Fig. 1

From: 27th Annual Computational Neuroscience Meeting (CNS*2018): Part Two

Fig. 1

Convolutional neuronal networks with extra-classical receptive fields (CNNEx). (a) Schematic of network architectures with and without extra-classical receptive fields, termed CNNEx and CNN respectively. The classical receptive field is shown in blue and the extra-classical receptive field is shown in green. In CNNEx, neurons from the extra-classical receptive field can modulate the activity of other neurons via recurrent lateral connections learned in an unsupervised manner. (b) The MNIST and CIFAR-10 datasets used to evaluate our models. Example images, along with images corrupted with additive white gaussian noise (AWGN) and salt-and-pepper noise (SP) are shown. The tables below each set of images correspond to model performance (percent accuracy) on the different image sets

Back to article page