Multi-Digit Recognition Using a Space Displacement Neural Network

By O. Matan, C. Burges, Yann LeCun, John S. Denker. 1991

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Fully convolutional networks for semantic segmentation (2014, methodology): “pace of FCNs and relate prior models both historical and recent. Fully convolutional networks To our knowledge, the idea of extending a convnet to arbitrary-sized inputs ﬁrst appeared in Matan et al. [20], which extended the classic LeNet [21] to recognize strings of digits. Because their net was li”
Rethinking ImageNet Pre-Training (2018, background): “, [31, 38, 45])—a fact that is somewhat overlooked today.”
Instance-Sensitive Fully Convolutional Networks (2016, background): “The general concept of fully convolutional models dates back to at least two decades ago [13].”
Gradient-based learning applied to document recognition (1998, methodology): “…the input image, and to rely on the \character spotting" property of the recognizer, i.e. its ability to correctly recognize a well-centered character in its input (cid:12)eld, even in the presence of other characters besides it, while rejecting images containing no centered characters [26], [27].”
A survey on deep learning-based fine-grained object classification and semantic segmentation (2017, background): “volutional network([27]), is trained to generate dense output feature maps for locating object parts.”
The End is Nigh: Generic Solving of Text-based CAPTCHAs (2014, background): “Space displacement neural network that attempt to recognize digits without segmentation were introduced in [35].”
Part-Stacked CNN for Fine-Grained Visual Categorization (2015, methodology): “The contributions of this paper include: 1) we present a novel and efficient part-based CNN architecture for finegrained recognition; 2) our architecture adopts an FCN to localize object parts, which has seldom been studied before in the context of object recognition; 3) our classification network f”
Markovian Models for Sequential Data (2004, background): “Recently, HMMs have been applied to a variety of applications outside of speech recognition, such as handwriting recognition [25, 26, 27, 28, 29, 30, 31], pattern recognition in molecular biology [32, 33, 34, 35, 3], and fault-detection [36].”
Object Recognition with Gradient-Based Learning (1999, methodology): “Although the idea of SDNN is quite old [10, 22 ], and very attractive by its simplicity, it has not generated wide interest until recently because of the enormous demands it puts on the recognizer.”
Deep Shape from a Low Number of Silhouettes (2016, background): “Also, fully convolutional networks prove their eﬃciency in one dimensional input strings [12] extended from LeNet [13], two dimensional detection [14,15] with learning and inference, three dimensional representation [16], volumetric 3D shapes classiﬁcation and interpolation [17].”
STFCN: Spatio-Temporal FCN for Semantic Video Segmentation (2016, methodology): “On the way of its development, it has been used for 1-dimensional (1D) and 2-dimensional (2D) inputs [34,44], and for solving various tasks such as image restoration, sliding window detection, depth estimation, boundary prediction, or semantic segmentation.”
Text-based CAPTCHA strengths and weaknesses (2011, background): “The space displacement neural network that attempts to recognize digits without segmentation was introduced in [24].”
PixelNet: Towards a General Pixel-level Architecture (2016, methodology): “Neural networks with convolutional output predictions, also called Fully Convolutional Networks (FCNs) [13, 48, 51, 57], appear to be a promising architecture in this direction.”