A Bayesian Approach to Image Recognition Based on Separable Lattice Hidden Markov Models

Kei SAWADA; Akira TAMAMORI; Kei HASHIMOTO; Yoshihiko NANKAKU; Keiichi TOKUDA

doi:10.1587/transinf.2016EDP7112

IEICE TRANSACTIONS on Information

A Bayesian Approach to Image Recognition Based on Separable Lattice Hidden Markov Models

Kei SAWADA, Akira TAMAMORI, Kei HASHIMOTO, Yoshihiko NANKAKU, Keiichi TOKUDA

Full Text Views

0

Cite this

Summary :

This paper proposes a Bayesian approach to image recognition based on separable lattice hidden Markov models (SL-HMMs). The geometric variations of the object to be recognized, e.g., size, location, and rotation, are an essential problem in image recognition. SL-HMMs, which have been proposed to reduce the effect of geometric variations, can perform elastic matching both horizontally and vertically. This makes it possible to model not only invariances to the size and location of the object but also nonlinear warping in both dimensions. The maximum likelihood (ML) method has been used in training SL-HMMs. However, in some image recognition tasks, it is difficult to acquire sufficient training data, and the ML method suffers from the over-fitting problem when there is insufficient training data. This study aims to accurately estimate SL-HMMs using the maximum a posteriori (MAP) and variational Bayesian (VB) methods. The MAP and VB methods can utilize prior distributions representing useful prior information, and the VB method is expected to obtain high generalization ability by marginalization of model parameters. Furthermore, to overcome the local maximum problem in the MAP and VB methods, the deterministic annealing expectation maximization algorithm is applied for training SL-HMMs. Face recognition experiments performed on the XM2VTS database indicated that the proposed method offers significantly improved image recognition performance. Additionally, comparative experiment results showed that the proposed method was more robust to geometric variations than convolutional neural networks.

Publication: IEICE TRANSACTIONS on Information Vol.E99-D No.12 pp.3119-3131

Publication Date: 2016/12/01

Publicized: 2016/09/05

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2016EDP7112

Type of Manuscript: PAPER

Category: Pattern Recognition

Authors

Kei SAWADA
  Nagoya Institute of Technology
Akira TAMAMORI
  Nagoya Institute of Technology
Kei HASHIMOTO
  Nagoya Institute of Technology
Yoshihiko NANKAKU
  Nagoya Institute of Technology
Keiichi TOKUDA
  Nagoya Institute of Technology

Keyword

image recognition, hidden Markov models, separable lattice hidden Markov models, Bayesian approach, deterministic annealing

Cite this

Copy

Kei SAWADA, Akira TAMAMORI, Kei HASHIMOTO, Yoshihiko NANKAKU, Keiichi TOKUDA, "A Bayesian Approach to Image Recognition Based on Separable Lattice Hidden Markov Models" in IEICE TRANSACTIONS on Information, vol. E99-D, no. 12, pp. 3119-3131, December 2016, doi: 10.1587/transinf.2016EDP7112.
Abstract: This paper proposes a Bayesian approach to image recognition based on separable lattice hidden Markov models (SL-HMMs). The geometric variations of the object to be recognized, e.g., size, location, and rotation, are an essential problem in image recognition. SL-HMMs, which have been proposed to reduce the effect of geometric variations, can perform elastic matching both horizontally and vertically. This makes it possible to model not only invariances to the size and location of the object but also nonlinear warping in both dimensions. The maximum likelihood (ML) method has been used in training SL-HMMs. However, in some image recognition tasks, it is difficult to acquire sufficient training data, and the ML method suffers from the over-fitting problem when there is insufficient training data. This study aims to accurately estimate SL-HMMs using the maximum a posteriori (MAP) and variational Bayesian (VB) methods. The MAP and VB methods can utilize prior distributions representing useful prior information, and the VB method is expected to obtain high generalization ability by marginalization of model parameters. Furthermore, to overcome the local maximum problem in the MAP and VB methods, the deterministic annealing expectation maximization algorithm is applied for training SL-HMMs. Face recognition experiments performed on the XM2VTS database indicated that the proposed method offers significantly improved image recognition performance. Additionally, comparative experiment results showed that the proposed method was more robust to geometric variations than convolutional neural networks.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2016EDP7112/_p

Copy

@ARTICLE{e99-d_12_3119,
author={Kei SAWADA, Akira TAMAMORI, Kei HASHIMOTO, Yoshihiko NANKAKU, Keiichi TOKUDA, },
journal={IEICE TRANSACTIONS on Information},
title={A Bayesian Approach to Image Recognition Based on Separable Lattice Hidden Markov Models},
year={2016},
volume={E99-D},
number={12},
pages={3119-3131},
abstract={This paper proposes a Bayesian approach to image recognition based on separable lattice hidden Markov models (SL-HMMs). The geometric variations of the object to be recognized, e.g., size, location, and rotation, are an essential problem in image recognition. SL-HMMs, which have been proposed to reduce the effect of geometric variations, can perform elastic matching both horizontally and vertically. This makes it possible to model not only invariances to the size and location of the object but also nonlinear warping in both dimensions. The maximum likelihood (ML) method has been used in training SL-HMMs. However, in some image recognition tasks, it is difficult to acquire sufficient training data, and the ML method suffers from the over-fitting problem when there is insufficient training data. This study aims to accurately estimate SL-HMMs using the maximum a posteriori (MAP) and variational Bayesian (VB) methods. The MAP and VB methods can utilize prior distributions representing useful prior information, and the VB method is expected to obtain high generalization ability by marginalization of model parameters. Furthermore, to overcome the local maximum problem in the MAP and VB methods, the deterministic annealing expectation maximization algorithm is applied for training SL-HMMs. Face recognition experiments performed on the XM2VTS database indicated that the proposed method offers significantly improved image recognition performance. Additionally, comparative experiment results showed that the proposed method was more robust to geometric variations than convolutional neural networks.},
keywords={},
doi={10.1587/transinf.2016EDP7112},
ISSN={1745-1361},
month={December},}

Copy

TY - JOUR
TI - A Bayesian Approach to Image Recognition Based on Separable Lattice Hidden Markov Models
T2 - IEICE TRANSACTIONS on Information
SP - 3119
EP - 3131
AU - Kei SAWADA
AU - Akira TAMAMORI
AU - Kei HASHIMOTO
AU - Yoshihiko NANKAKU
AU - Keiichi TOKUDA
PY - 2016
DO - 10.1587/transinf.2016EDP7112
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E99-D
IS - 12
JA - IEICE TRANSACTIONS on Information
Y1 - December 2016
AB - This paper proposes a Bayesian approach to image recognition based on separable lattice hidden Markov models (SL-HMMs). The geometric variations of the object to be recognized, e.g., size, location, and rotation, are an essential problem in image recognition. SL-HMMs, which have been proposed to reduce the effect of geometric variations, can perform elastic matching both horizontally and vertically. This makes it possible to model not only invariances to the size and location of the object but also nonlinear warping in both dimensions. The maximum likelihood (ML) method has been used in training SL-HMMs. However, in some image recognition tasks, it is difficult to acquire sufficient training data, and the ML method suffers from the over-fitting problem when there is insufficient training data. This study aims to accurately estimate SL-HMMs using the maximum a posteriori (MAP) and variational Bayesian (VB) methods. The MAP and VB methods can utilize prior distributions representing useful prior information, and the VB method is expected to obtain high generalization ability by marginalization of model parameters. Furthermore, to overcome the local maximum problem in the MAP and VB methods, the deterministic annealing expectation maximization algorithm is applied for training SL-HMMs. Face recognition experiments performed on the XM2VTS database indicated that the proposed method offers significantly improved image recognition performance. Additionally, comparative experiment results showed that the proposed method was more robust to geometric variations than convolutional neural networks.
ER -

IEICE TRANSACTIONS on Information