Work at the intersection of vision science and deep learning is starting to explore the efficacy of deep convolutional networks (DCNs) and recurrent networks in solving perceptual grouping problems that underlie primate visual recognition and segmentation . We introduce V1Net, a bio-inspired recurrent unit that incorporates lateral connections ubiquitous in cortical circuitry . We compare the learning efficiency and accuracy of V1net-DCNs to that of 14 carefully selected feedforward and recurrent neural architectures (including state-of-the-art DCNs) on MarkedLong — a synthetic forced-choice contour integration dataset of 800,000 images we introduce here . We also note that V1NET-DCN learns the most compact generalizable solution to Marked long . The compact and rich representations make it a promising candidate to build on-device machine vision algorithms as well as help better understand biological cortical circuitry and help better understanding biological cortical circuitry algorithms as well as machine vision algorithm asphyphyphytastic technically technique and deep-learning techniques . We gauged solution generalizability by measuring the transfer learning performance of our candidate models trained on PathFinder. To learn more, please contact us at [ .

Links: PDF - Abstract

Code :


Keywords : learning - cortical - vision - circuitry - deep -

Leave a Reply

Your email address will not be published. Required fields are marked *


Enjoy this blog? Please spread the word :)