Recurrent neural circuits for contour detection

We introduce a deep recurrent neural network architecture that approximates visual cortical circuits (Mely et al., 2018). We show that this architecture, which we refer to as the 𝜸-net, learns to solve contour detection tasks with better sample efficiency than state-of-the-art feedforward networks, while also exhibiting a classic perceptual illusion, known as the orientation-tilt illusion. Correcting this illusion significantly reduces \gnetw contour detection accuracy by driving it to prefer low-level edges over high-level object boundary contours. Overall, our study suggests that the orientation-tilt illusion is a byproduct of neural circuits that help biological visual systems achieve robust and efficient contour detection, and that incorporating these circuits in artificial neural networks can improve computer vision.

[1]  Richard Hans Robert Hahnloser,et al.  Digital selection and analogue amplification coexist in a cortex-inspired silicon circuit , 2000, Nature.

[2]  Charless C. Fowlkes,et al.  Contour Detection and Hierarchical Image Segmentation , 2011, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[3]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[4]  Charless C. Fowlkes,et al.  Oriented edge forests for boundary detection , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[5]  Joshua B. Tenenbaum,et al.  Human-level concept learning through probabilistic program induction , 2015, Science.

[6]  Aaron C. Courville,et al.  Recurrent Batch Normalization , 2016, ICLR.

[7]  Nikolaus Kriegeskorte,et al.  Deep neural networks: a new framework for modelling biological vision and brain information processing , 2015, bioRxiv.

[8]  William R. Gray Roncal,et al.  Saturated Reconstruction of a Volume of Neocortex , 2015, Cell.

[9]  Selmaan N. Chettih,et al.  Single-neuron perturbations reveal feature-specific competition in V1 , 2019, Nature.

[10]  Nikolaus Kriegeskorte,et al.  Recurrence is required to capture the representational dynamics of the human visual system , 2019, Proceedings of the National Academy of Sciences.

[11]  Xiang Bai,et al.  Richer Convolutional Features for Edge Detection , 2016, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[12]  Thomas Serre,et al.  Learning long-range spatial dependencies with horizontal gated-recurrent units , 2018, NeurIPS.

[13]  David J. Jilk,et al.  Recurrent Processing during Object Recognition , 2011, Front. Psychol..

[14]  David Cox,et al.  Recurrent computations for visual pattern completion , 2017, Proceedings of the National Academy of Sciences.

[15]  Gabriel Kreiman,et al.  Deep Predictive Coding Networks for Video Prediction and Unsupervised Learning , 2016, ICLR.

[16]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[17]  Tai Sing Lee,et al.  Hierarchical Bayesian inference in the visual cortex. , 2003, Journal of the Optical Society of America. A, Optics, image science, and vision.

[18]  Christopher Joseph Pal,et al.  On orthogonality and learning recurrent networks with long term dependencies , 2017, ICML.

[19]  Benjamin Recht,et al.  Do CIFAR-10 Classifiers Generalize to CIFAR-10? , 2018, ArXiv.

[20]  Thomas Serre,et al.  Not-So-CLEVR: learning same–different relations strains feedforward neural networks , 2018, Interface Focus.

[21]  Iasonas Kokkinos,et al.  Pushing the Boundaries of Boundary Detection using Deep Learning , 2015, ICLR 2016.

[22]  Alon Poleg-Polsky,et al.  Species-specific wiring for direction selectivity in the mammalian retina , 2016, Nature.

[23]  Thomas Serre,et al.  Learning what and where to attend , 2018, ICLR.

[24]  Ming Yang,et al.  Bi-Directional Cascade Network for Perceptual Edge Detection , 2019, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[25]  Michał Januszewski,et al.  Segmentation-Enhanced CycleGAN , 2019, bioRxiv.

[26]  J. DiCarlo,et al.  Using goal-driven deep learning models to understand sensory cortex , 2016, Nature Neuroscience.

[27]  Markus Siegel,et al.  Cortical information flow during flexible sensorimotor decisions , 2015, Science.

[28]  Shengjun Liu,et al.  Learning to predict crisp boundaries , 2018, ECCV.

[29]  Yann Ollivier,et al.  Can recurrent neural networks warp time? , 2018, ICLR.

[30]  Fathi M. Salem,et al.  Simplified minimal gated unit variations for recurrent neural networks , 2017, 2017 IEEE 60th International Midwest Symposium on Circuits and Systems (MWSCAS).

[31]  C. Gilbert,et al.  Brain States: Top-Down Influences in Sensory Processing , 2007, Neuron.

[32]  H. Sebastian Seung,et al.  Superhuman Accuracy on the SNEMI3D Connectomics Challenge , 2017, ArXiv.

[33]  Eugenio Culurciello,et al.  Deep Predictive Coding Network for Object Recognition , 2018, ICML.

[34]  Bhaskara Marthi,et al.  A generative vision model that trains with high data efficiency and breaks text-based CAPTCHAs , 2017, Science.

[35]  Swami Sankaranarayanan,et al.  Face recognition accuracy of forensic examiners, superrecognizers, and face recognition algorithms , 2018, Proceedings of the National Academy of Sciences.

[36]  Jianbo Shi,et al.  Machine Learning of Hierarchical Clustering to Segment 2D and 3D Images , 2013, PloS one.

[37]  Eric T. Shea-Brown,et al.  Dynamic representation of partially occluded objects in primate prefrontal and visual cortex , 2017, eLife.

[38]  Matthias Bethge,et al.  Generalisation in humans and deep neural networks , 2018, NeurIPS.

[39]  P. Roelfsema,et al.  Distinct Feedforward and Feedback Effects of Microstimulation in Visual Cortex Reveal Neural Mechanisms of Texture Segregation , 2017, Neuron.

[40]  Thomas Serre,et al.  Deep Learning: The Good, the Bad, and the Ugly. , 2019, Annual review of vision science.

[41]  Nikolaus Kriegeskorte,et al.  Recurrent Convolutional Neural Networks: A Better Model of Biological Object Recognition , 2017, bioRxiv.

[42]  Haluk Öğmen,et al.  Feedforward and feedback processes in vision , 2015, Front. Psychol..

[43]  Kaiqi Huang,et al.  Deep Crisp Boundaries: From Boundaries to Higher-Level Tasks , 2018, IEEE Transactions on Image Processing.

[44]  J. Dunning.  The elephant in the room. , 2013, European journal of cardio-thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery.

[45]  Sander W. Keemink,et al.  A unified account of tilt illusions, association fields, and contour detection based on elastica , 2016, Vision Research.

[46]  Davi D Bock,et al.  Volume electron microscopy for neuronal circuit reconstruction , 2012, Current Opinion in Neurobiology.

[47]  Thomas Serre,et al.  Complementary Surrounds Explain Diverse Contextual Phenomena Across Visual Modalities , 2018, Psychological review.

[48]  David J. Jilk,et al.  Early recurrent feedback facilitates visual object recognition under challenging conditions , 2014, Front. Psychol..

[49]  Thomas Serre,et al.  Robust neural circuit reconstruction from serial electron microscopy with convolutional recurrent networks , 2018, ArXiv.

[50]  Peter Wenderoth,et al.  The tilt illusion: Repulsion and attraction effects in the oblique meridian , 1977, Vision Research.

[51]  James J. DiCarlo,et al.  Evidence that recurrent circuits are critical to the ventral stream’s execution of core object recognition behavior , 2018, Nature Neuroscience.

[52]  Tomaso A. Poggio,et al.  Bridging the Gaps Between Residual Learning, Recurrent Neural Networks and Visual Cortex , 2016, ArXiv.

[53]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[54]  S. Hochstein,et al.  View from the Top Hierarchies and Reverse Hierarchies in the Visual System , 2002, Neuron.

[55]  Michael C. Mozer,et al.  Induction of Multiscale Temporal Structure , 1991, NIPS.

[56]  Andrea Vedaldi,et al.  Instance Normalization: The Missing Ingredient for Fast Stylization , 2016, ArXiv.

[57]  Surya Ganguli,et al.  Task-Driven Convolutional Recurrent Models of the Visual System , 2018, NeurIPS.

[58]  P. Roelfsema.  Cortical algorithms for perceptual grouping. , 2006, Annual review of neuroscience.

[59]  D. J. Felleman,et al.  Distributed hierarchical processing in the primate cerebral cortex. , 1991, Cerebral cortex.