Signature Verification Using a “Siamese” Time Delay Neural Network

Bromley, Jane; Guyon, Isabelle; LeCun, Yann; Sackinger, E.; Shah, Roopak

doi:10.1142/9789812797926_0003

Cited by 2,051 publications

(1,901 citation statements)

References 4 publications

Supporting

Mentioning

1,881

Contrasting

Unclassified

Order By: Relevance

“…For this reason, we learn the relationship between feature pairs instead of individual features. Motivated by Siamese network [7], we use the label of feature pair to describe the relationship between individual features, and the relationship is learned by minimizing the following contrastive loss function:…”

Section: Relationship Representationmentioning

confidence: 99%

Learning Deep Relationship for Object Detection

Huo

2018

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYObject detection has been a hot topic of image processing, computer vision and pattern recognition. In recent years, training a model from labeled images using machine learning technique becomes popular. However, the relationship between training samples is usually ignored by existing approaches. To address this problem, a novel approach is proposed, which trains Siamese convolutional neural network on feature pairs and finely tunes the network driven by a small amount of training samples. Since the proposed method considers not only the discriminative information between objects and background, but also the relationship between intraclass features, it outperforms the state-of-arts on real images.

show abstract

Section: Relationship Representationmentioning

confidence: 99%

Learning Deep Relationship for Object Detection

Huo

2018

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

show abstract

“…This type of NN are valuables for class of problems in which the temporal information as well as the relationship between the input and output sequences are required. It has been applied in various applications, amongst the lots are signature verification [24], character recognition [25], and spoken word recognition [23]. …”

Section: B Neural Network Topologiesmentioning

confidence: 99%

On the Problem of Features Variability in Sequence Learning Problems

Yahaya¹

2015

IJFCC

View full text Add to dashboard Cite

Abstract-Sequential learning problems such as speech, cursive handwriting, time series forecasting and protein sequence prediction. Both Speech and cursive handwriting recognition are challenging problems to Pattern recognition systems, in particular speech signal. Some peculiar characteristics of these types of problems are that, the signal or pattern evolves with time, modeling a long time dependencies in this pattern is a major challenge. Hidden Markov models (HMM) have been applied for these types of problems. Due to some obvious shortcomings of HMM, neural networks were also explored and applied as well as their hybrids. The problem of feature variability in sequence learning is still a challenging problem. In this paper, we analyzed the problem, present some methods in feature variance suppression in character recognition, and review some research efforts in modification of neural networks and applications. We proposed a structure for a state-based neural network.Index Terms-Sequence learning, feature variability, neural network.

show abstract

“…Within the DrLim framework the same network is applied to two different inputs in order to compute the loss for this input pair (see equation 10). Therefore, the architecture is sometimes called a siamese network [12,13]. In this work we investigate two aspects of a configuration of a Convolutional Neural Network:…”

Section: Convolutional Neural Networkmentioning

confidence: 99%

Convolutional Neural Networks Learn Compact Local Image Descriptors

Osendorfer

Bayer

Urban

et al. 2013

Neural Information Processing

View full text Add to dashboard Cite

Abstract. We investigate if a deep Convolutional Neural Network can learn representations of local image patches that are usable in the important task of keypoint matching. We examine several possible loss functions for this correspondance task and show emprically that a newly suggested loss formulation allows a Convolutional Neural Network to find compact local image descriptors that perform comparably to stateof-the-art approaches.

show abstract

Signature Verification Using a “Siamese” Time Delay Neural Network

Cited by 2,051 publications

References 4 publications

Learning Deep Relationship for Object Detection

Learning Deep Relationship for Object Detection

On the Problem of Features Variability in Sequence Learning Problems

Convolutional Neural Networks Learn Compact Local Image Descriptors

Contact Info

Product

Resources

About