FSRFNet: Feature-Selective and Spatial Receptive Fields Networks

The attention mechanism plays a crucial role in the human visual experience. In the cognitive neuroscience community, the receptive field size of visual cortical neurons is regulated by the additive effect of feature-selective and spatial attention. We propose a novel architectural unit called a “Feature-selective and Spatial Receptive Fields” (FSRF) block that implements adaptive receptive field sizes of neurons through the additive effects of feature-selective and spatial attention. We show that FSRF blocks can be inserted into the architecture of existing convolutional neural networks to form an FSRF network architecture, and test its generalization capabilities on different datasets.

[1]  Nikos Komodakis,et al.  Wide Residual Networks , 2016, BMVC.

[2]  Jian Sun,et al.  Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition , 2015, IEEE Trans. Pattern Anal. Mach. Intell..

[3]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[4]  D. Hubel,et al.  Receptive fields, binocular interaction and functional architecture in the cat's visual cortex , 1962, The Journal of physiology.

[5]  Christof Koch,et al.  A Model of Saliency-Based Visual Attention for Rapid Scene Analysis , 2009 .

[6]  R. Desimone,et al.  Attentional control of visual perception: cortical and subcortical mechanisms. , 1990, Cold Spring Harbor symposia on quantitative biology.

[7]  C. Koch,et al.  Computational modelling of visual attention , 2001, Nature Reviews Neuroscience.

[8]  Qiang Chen,et al.  Network In Network , 2013, ICLR.

[9]  J. Nelson,et al.  Orientation-selective inhibition from beyond the classic visual receptive field , 1978, Brain Research.

[10]  Jan Matuszewski,et al.  Radar signal identification using a neural network and pattern recognition methods , 2018, 2018 14th International Conference on Advanced Trends in Radioelecrtronics, Telecommunications and Computer Engineering (TCSET).

[11]  Jan Matuszewski,et al.  Recognition of electromagnetic sources with the use of deep neural networks , 2019, Conference on Reconnaissance and Electronic Warfare Systems.

[12]  Bo Chen,et al.  MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications , 2017, ArXiv.

[13]  Søren K. Andersen,et al.  Effects of Feature-selective and Spatial Attention at Different Stages of Visual Processing , 2011, Journal of Cognitive Neuroscience.

[14]  Jim M. Monti,et al.  Neural Integration of Top-Down Spatial and Feature-Based Information in Visual Search , 2008, The Journal of Neuroscience.

[15]  Janusz Dudczyk,et al.  Radar Emission Sources Identification Based on Hierarchical Agglomerative Clustering for Large Data Sets , 2016, J. Sensors.

[16]  Karl J. Friston,et al.  The Relationship Between Synchronization Among Neuronal Populations and Their Mean Activity Levels , 1999, Neural Computation.

[17]  Koen E. A. van de Sande,et al.  Selective Search for Object Recognition , 2013, International Journal of Computer Vision.

[18]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[19]  David J. Freedman,et al.  Interaction between Spatial and Feature Attention in Posterior Parietal Cortex , 2016, Neuron.

[20]  R. Desimone,et al.  Neural mechanisms of spatial selective attention in areas V1, V2, and V4 of macaque visual cortex. , 1997, Journal of neurophysiology.

[21]  Leslie G. Ungerleider,et al.  Mechanisms of visual attention in the human cortex. , 2000, Annual review of neuroscience.

[22]  Yoshua Bengio,et al.  Neural Machine Translation by Jointly Learning to Align and Translate , 2014, ICLR.

[23]  J. Gallant,et al.  Time Course of Attention Reveals Different Mechanisms for Spatial and Feature-Based Attention in Area V4 , 2005, Neuron.

[24]  Stanford,et al.  Tiny ImageNet Classification with Convolutional Neural Networks , 2015 .

[25]  Hendrik Strumpf,et al.  Enhanced spatial focusing increases feature-based selection in unattended locations , 2018, Scientific Reports.

[26]  B. C. Motter.  Focal attention produces spatially selective processing in visual cortical areas V1, V2, and V4 in the presence of competing stimuli. , 1993, Journal of neurophysiology.