Accurate Thickness Measurement of Two Adjacent Sheet Structures in CT Images

Yuanzhi CHENG; Yoshinobu SATO; Hisashi TANAKA; Takashi NISHII; Nobuhiko SUGANO; Hironobu NAKAMURA; Hideki YOSHIKAWA; Shuguo WANG; Shinichi TAMURA

Accurate Thickness Measurement of Two Adjacent Sheet Structures in CT Images

Yuanzhi CHENG, Yoshinobu SATO, Hisashi TANAKA, Takashi NISHII, Nobuhiko SUGANO, Hironobu NAKAMURA, Hideki YOSHIKAWA, Shuguo WANG, Shinichi TAMURA

Full Text Views

0

Cite this

Summary :

Accurate thickness measurement of sheet-like structure such as articular cartilage in CT images is required in clinical diagnosis as well as in fundamental research. Using a conventional measurement method based on the zero-crossing edge detection (zero-crossings method), several studies have already analyzed the accuracy limitation on thickness measurement of the single sheet structure that is not influenced by peripheral structures. However, no studies, as of yet, have assessed measurement accuracy of two adjacent sheet structures such as femoral and acetabular cartilages in the hip joint. In this paper, we present a model of the CT scanning process of two parallel sheet structures separated by a small distance, and use the model to predict the shape of the gray-level profiles along the sheet normal orientation. The difference between the predicted and the actual gray-level profiles observed in the CT data is minimized by refining the model parameters. Both a one-by-one search (exhaustive combination search) technique and a nonlinear optimization technique based on the Levenberg-Marquardt algorithm are used to minimize the difference. Using CT images of phantoms, we present results showing that when applying the one-by-one search method to obtain the initial values of the model parameters, Levenberg-Marquardt method is more accurate than zero-crossings and one-by-one search methods for estimating the thickness of two adjacent sheet structures, as well as the thickness of a single sheet structure.

Publication: IEICE TRANSACTIONS on Information Vol.E90-D No.1 pp.271-282

Publication Date: 2007/01/01

Publicized

Online ISSN: 1745-1361

DOI

Type of Manuscript: Special Section PAPER (Special Section on Advanced Image Technology)

Category

Cite this

Copy

Yuanzhi CHENG, Yoshinobu SATO, Hisashi TANAKA, Takashi NISHII, Nobuhiko SUGANO, Hironobu NAKAMURA, Hideki YOSHIKAWA, Shuguo WANG, Shinichi TAMURA, "Accurate Thickness Measurement of Two Adjacent Sheet Structures in CT Images" in IEICE TRANSACTIONS on Information, vol. E90-D, no. 1, pp. 271-282, January 2007, doi: .
Abstract: Accurate thickness measurement of sheet-like structure such as articular cartilage in CT images is required in clinical diagnosis as well as in fundamental research. Using a conventional measurement method based on the zero-crossing edge detection (zero-crossings method), several studies have already analyzed the accuracy limitation on thickness measurement of the single sheet structure that is not influenced by peripheral structures. However, no studies, as of yet, have assessed measurement accuracy of two adjacent sheet structures such as femoral and acetabular cartilages in the hip joint. In this paper, we present a model of the CT scanning process of two parallel sheet structures separated by a small distance, and use the model to predict the shape of the gray-level profiles along the sheet normal orientation. The difference between the predicted and the actual gray-level profiles observed in the CT data is minimized by refining the model parameters. Both a one-by-one search (exhaustive combination search) technique and a nonlinear optimization technique based on the Levenberg-Marquardt algorithm are used to minimize the difference. Using CT images of phantoms, we present results showing that when applying the one-by-one search method to obtain the initial values of the model parameters, Levenberg-Marquardt method is more accurate than zero-crossings and one-by-one search methods for estimating the thickness of two adjacent sheet structures, as well as the thickness of a single sheet structure.
URL: https://global.ieice.org/en_transactions/information/10.1587/e90-d_1_271/_p

Copy

@ARTICLE{e90-d_1_271,
author={Yuanzhi CHENG, Yoshinobu SATO, Hisashi TANAKA, Takashi NISHII, Nobuhiko SUGANO, Hironobu NAKAMURA, Hideki YOSHIKAWA, Shuguo WANG, Shinichi TAMURA, },
journal={IEICE TRANSACTIONS on Information},
title={Accurate Thickness Measurement of Two Adjacent Sheet Structures in CT Images},
year={2007},
volume={E90-D},
number={1},
pages={271-282},
abstract={Accurate thickness measurement of sheet-like structure such as articular cartilage in CT images is required in clinical diagnosis as well as in fundamental research. Using a conventional measurement method based on the zero-crossing edge detection (zero-crossings method), several studies have already analyzed the accuracy limitation on thickness measurement of the single sheet structure that is not influenced by peripheral structures. However, no studies, as of yet, have assessed measurement accuracy of two adjacent sheet structures such as femoral and acetabular cartilages in the hip joint. In this paper, we present a model of the CT scanning process of two parallel sheet structures separated by a small distance, and use the model to predict the shape of the gray-level profiles along the sheet normal orientation. The difference between the predicted and the actual gray-level profiles observed in the CT data is minimized by refining the model parameters. Both a one-by-one search (exhaustive combination search) technique and a nonlinear optimization technique based on the Levenberg-Marquardt algorithm are used to minimize the difference. Using CT images of phantoms, we present results showing that when applying the one-by-one search method to obtain the initial values of the model parameters, Levenberg-Marquardt method is more accurate than zero-crossings and one-by-one search methods for estimating the thickness of two adjacent sheet structures, as well as the thickness of a single sheet structure.},
keywords={},
doi={},
ISSN={1745-1361},
month={January},}

Copy

TY - JOUR
TI - Accurate Thickness Measurement of Two Adjacent Sheet Structures in CT Images
T2 - IEICE TRANSACTIONS on Information
SP - 271
EP - 282
AU - Yuanzhi CHENG
AU - Yoshinobu SATO
AU - Hisashi TANAKA
AU - Takashi NISHII
AU - Nobuhiko SUGANO
AU - Hironobu NAKAMURA
AU - Hideki YOSHIKAWA
AU - Shuguo WANG
AU - Shinichi TAMURA
PY - 2007
DO -
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E90-D
IS - 1
JA - IEICE TRANSACTIONS on Information
Y1 - January 2007
AB - Accurate thickness measurement of sheet-like structure such as articular cartilage in CT images is required in clinical diagnosis as well as in fundamental research. Using a conventional measurement method based on the zero-crossing edge detection (zero-crossings method), several studies have already analyzed the accuracy limitation on thickness measurement of the single sheet structure that is not influenced by peripheral structures. However, no studies, as of yet, have assessed measurement accuracy of two adjacent sheet structures such as femoral and acetabular cartilages in the hip joint. In this paper, we present a model of the CT scanning process of two parallel sheet structures separated by a small distance, and use the model to predict the shape of the gray-level profiles along the sheet normal orientation. The difference between the predicted and the actual gray-level profiles observed in the CT data is minimized by refining the model parameters. Both a one-by-one search (exhaustive combination search) technique and a nonlinear optimization technique based on the Levenberg-Marquardt algorithm are used to minimize the difference. Using CT images of phantoms, we present results showing that when applying the one-by-one search method to obtain the initial values of the model parameters, Levenberg-Marquardt method is more accurate than zero-crossings and one-by-one search methods for estimating the thickness of two adjacent sheet structures, as well as the thickness of a single sheet structure.
ER -