Stiffness prediction on elastography images and neuro-fuzzy based segmentation for thyroid cancer detection

Koushik Layek; Biswanath Basak; Sourav Samanta; Santi Prasad Maity; Ananya Barui

doi:10.1364/AO.445226

Applied Optics
Vol. 61,
Issue 1,
pp. 49-59
(2022)
•https://doi.org/10.1364/AO.445226

Stiffness prediction on elastography images and neuro-fuzzy based segmentation for thyroid cancer detection

Koushik Layek, Biswanath Basak, Sourav Samanta, Santi Prasad Maity, and Ananya Barui

Not Accessible

Your library or personal account may give you access

Get PDF
Email
Share
Get Citation
Copy Citation Text
Koushik Layek, Biswanath Basak, Sourav Samanta, Santi Prasad Maity, and Ananya Barui, "Stiffness prediction on elastography images and neuro-fuzzy based segmentation for thyroid cancer detection," Appl. Opt. 61, 49-59 (2022)

Export Citation
- BibTex
- Endnote (RIS)
- HTML
- Plain Text
Citation alert
Save article
Editors' Pick

Check for updates

Related Topics
Table of Contents Category
- Imaging Systems, Image Processing, and Displays
Optics & Photonics Topics
?

The topics in this list come from the Optics and Photonics Topics applied to this article.

About this Article
History
- Original Manuscript: October 5, 2021
- Revised Manuscript: November 23, 2021
- Manuscript Accepted: November 29, 2021
- Published: December 21, 2021

Abstract

The elastography method detects metastatic changes by measuring the stiffness of tissues. Estimation of elasticities from elastography images facilitates more precise identification of the metastatic region and detection of the same. In this study, an automated segmentation algorithm is proposed that calculates pixel-wise elasticity values to detect thyroid cancer from elastography images. This intensity to elasticity conversion is achieved by constructing a fuzzy inference system using an adaptive neuro-fuzzy inference system supported by two meta-heuristic algorithms: genetic algorithm and particle swarm optimization. Pixels of the input color images (red, green, and blue) are replaced by equivalent elasticity values (in kilo Pascal) and are stored in a two-dimensional array to form an “elasticity matrix.” The elasticity matrix is then segmented into three regions, namely, suspicious, near-suspicious, and non-suspicious, based on the elasticity measures, where the threshold limits are calculated using the fuzzy entropy maximization method optimized by the differential evolution algorithm. Segmentation performances are evaluated by Kappa and the dice similarity co-efficient, and average values achieved are $0.94 \pm 0.11$ and $0.93 \pm 0.12$, respectively. Sensitivity and specificity values achieved by the proposed method are $86.35 \pm 0.34\%$ and $97.67 \pm 0.40\%$, respectively, showing an overall accuracy of $93.50 \pm 0.42\%$. Results justify the importance of pixel stiffness for segmentation of thyroid nodules in elastography images.

Full Article | PDF Article

More Like This

Photoacoustic/ultrasound dual imaging of human thyroid cancers: an initial clinical study

Meng Yang, Lingyi Zhao, Xujin He, Na Su, ChenYang Zhao, Hewen Tang, Tao Hong, Wenbo Li, Fang Yang, Lin Lin, Bing Zhang, Rui Zhang, Yuxin Jiang, and Changhui Li
Biomed. Opt. Express 8(7) 3449-3457 (2017)

Remote sensing image segmentation based on a robust fuzzy C-means algorithm improved by a parallel Lévy grey wolf algorithm

M. Q. Li, L. P. Xu, Shan Gao, Na Xu, and Bo Yan
Appl. Opt. 58(17) 4812-4822 (2019)

Adaptive neuro fuzzy selection of important factors for prediction of plasmons in silver nanorods

Dalibor Petković, Mohamed Amine Khadimallah, Yan Cao, Nebojsa Denic, Vuk Vujovic, Dragan Zlatkovic, and Jelena Stojanovic
Appl. Opt. 61(10) 2864-2868 (2022)

Previous Article Next Article

Supplementary Material (1)

Name	Description
Supplement 1	The supplementary document has been prepared so that additional information about the work can be provided with better clarity.

Data Availability

Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors and consent from the data collection center upon reasonable request. No data were generated or analyzed in the presented research. The data used are collected from EKO X Ray and Imaging Centre, Kolkata.

Cited By

You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Figures (8)

You do not have subscription access to this journal. Figure files are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Tables (6)

You do not have subscription access to this journal. Article tables are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Equations (23)

You do not have subscription access to this journal. Equations are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Literature	Algorithms Used	Performance Metrics	Limitations
Huang et al. [22] (2019)	AFGC, morphological features Dataset size: 610	SA: 0.9974 CSc: 0.9969	Window based computation, high computation required, not suitable for images with higher dimension
Buda et al. [23] (2020)	CNN Dataset used: 1278	Dice metric: 93.1%	Depends on manually marked calipers
Yang et al. [25] (2021)	U-Net, Ds-GAN, SVM, dataset	Accuracy: 90.01%	Several cascaded models used, cross-marks by expert required
		AUC: 91.07%
	Size: 3090	Sensitivity: 87.47%
	Size: 3090	Specificity: 92.15%
Proposed model	ANFIS, GA/PSO, FEM, DE	$κ$ : $0.94 \pm 0.11$ ,	Requires elasticity data with annotations, requires training to generate FIS
		DSC: $0.93 \pm 0.12$
		Sensitivity: $86.35 \pm 0.34 %$
		Specificity: $97.67 \pm 0.40 %$

Cross Validation	ANFIS	ANFIS-GA	ANFIS-PSO
3-fold	3.46	1.87	1.63
5-fold	3.44	1.83	1.61
7-fold	3.48	1.85	1.64
10-fold	3.51	1.86	1.63

	Red	Red	Green	Green	Blue	Blue
Methods	SSIM	PSNR (dB)	SSIM	PSNR (dB)	SSIM	PSNR (dB)
CS	0.61	15.36	0.76	26.54	0.2	21.98
HE	0.51	14.62	0.38	10.22	0.18	7.39
AHE	0.74	23.64	0.6	28.23	0.52	15.54
CLAHE	0.82	30.32	0.87	33.65	0.78	27.28

	Red	Red	Green	Green	Blue	Blue
Methods	SSIM	PSNR (dB)	SSIM	PSNR (dB)	SSIM	PSNR (dB)
CS	0.02	2.39	0.06	4.3	0.08	4.23
HE	0.03	2.16	0.05	1.89	0.04	1.64
AHE	0.05	3.70	0.07	3.83	0.07	2.73
CLAHE	0.07	4.61	0.12	4.52	0.12	5.29

Metrics	FEM-DE	TEM-DE	FCM	Otsu’s
$κ$	$0.94 \pm 0.11$	$0.89 \pm 0.65$	$0.85 \pm 0.52$	$0.77 \pm 0.16$
DSC	$0.93 \pm 0.12$	$0.87 \pm 0.23$	$0.84 \pm 0.17$	$0.75 \pm 0.38$
Sensitivity	$86.35 \pm 0.34 %$	$80.16 \pm 0.77 %$	$78.83 \pm 0.61 %$	$72.58 \pm 0.33 %$
Specificity	$97.67 \pm 0.40 %$	$94.13 \pm 0.29 %$	$91.29 \pm 0.94 %$	$79.22 \pm 0.52 %$
Accuracy	$93.50 \pm 0.42 %$	$89.50 \pm 0.05 %$	$87.13 \pm 0.59 %$	$76.91 \pm 0.16 %$

Metrics	FEM-DE	TEM-DE	FCM	Otsu’s
$κ$	$0.82 \pm 0.25$	$0.79 \pm 0.57$	$0.77 \pm 0.14$	$0.61 \pm 0.29$
DSC	$0.76 \pm 0.40$	$0.74 \pm 0.98$	$0.72 \pm 0.33$	$0.63 \pm 0.66$
Sensitivity	$85.5 \pm 0.42 %$	$81.28 \pm 0.21 %$	$79.06 \pm 0.16 %$	$69.10 \pm 0.20 %$
Specificity	$95.67 \pm 0.13 %$	$93.69 \pm 0.72 %$	$91.92 \pm 0.51 %$	$76.52 \pm 0.13 %$
Accuracy	$89.74 \pm 0.81 %$	$87.43 \pm 0.75 %$	$85.16 \pm 0.36 %$	$72.50 \pm 0.18 %$

Abstract

Supplementary Material (1)

Data Availability

Cited By

Figures (8)

Tables (6)

Equations (23)

Applied Optics