Abstract
Convolutional neural networks (CNNs) have played a significant role in feature extraction and tasks thereafter for accurate and automated diagnosis from ultrasound (US) breast tumor images. However, using pre-trained architectures and transfer learning for feature extraction could cause negative transfer in medical domain. Also publicly accessible online training/ validation US breast tumor datasets are seldom available. Hence, it becomes prudent to develop alternate CNN architectures as feature extraction backbones which are trained on smaller datasets without any consequences of overfitting. In this paper, a CNN was developed for feature extraction and prediction of breast tumor as benign/ malignant, with hyper parameters (learning rate, regularization factor, momentum, section depth and number of convolution filters) optimized using bayesian optimization. To further prevent any overfitting due to limited training data, a novel neutrosophic augmentation method was also introduced. The obtained simulation results on three different test datasets show that the classification accuracy of the optimized CNN outperforms by at least 3% than many other state-of-the-art deep architectures and significantly greater than 5% for shallow architectures. For segmenting the tumor region, the convolution maps from the higher layers of the optimized CNN are clustered to provide the initial contour for segmentation using active contours. The segmentation metrics with respect to ground truth is greater for the proposed work when compared with U-Net and fully Convolutional Network based segmentation. The structural similarity and mean segmentation error for the proposed method and the latter cases are 0.98, 0.93 and 0.92, and 0.01, 0.21 and 0.28 respectively.
Similar content being viewed by others
Data availability
Not provided.
Code availability
Not provided.
References
Al-Dhabyani W, Gomaa M, Khaled H, Fahmy A (2020) Dataset of breast ultrasound images. Data Brief 28:104863
Bergstra J, Bengio Y (2012) Random search for hyper-parameter optimization. J Mach Learn Res 13(2):281–305
Borgli RJ, Stensland HK, Riegler MA, Halvorsen P (2019) Automatic hyperparameter optimization for transfer learning on medical image datasets using Bayesian optimization. 2019 13th international symposium on medical information and communication technology. IEEE, pp 1–6
Bose A, Nguyen T, Du H, AlZoubi A (2021) Faster RCNN hyperparameter selection for breast lesion detection in 2D ultrasound images. In: UK workshop on computational intelligence, pp 179–190
Byra M, Jarosik P, Szubert A, Galperin M, Ojeda-Fournier H, Olson L, O’Boyle M, Comstock C, Andre M (2020) Breast mass segmentation in ultrasound with selective kernel U-Net convolutional neural network. Biomed Signal Process Control 61:102027
Ceylan Z (2020) Diagnosis of breast cancer using improved machine learning algorithms based on bayesian optimization. Int J Intell Syst Appl Eng 8(3):121–130
Chan TF, Vese LA (2001) Active contours without edges. IEEE Trans Image Process 10(2):266–277
Cheng HD, Shan J, Ju W, Guo Y, Zhang L (2010) Automated breast cancer detection and classification using ultrasound images: a survey. Pattern Recogn 43(1):299–317
Doke P, Shrivastava D, Pan C, Zhou Q, Zhang YD (2020) Using CNN with bayesian optimization to identify cerebral micro-bleeds. Mach Vis Appl 31:1–4
Han S, Kang HK, Jeong JY, Park MH, Kim W, Bang WC, Seong YK (2017) A deep learning framework for supporting the classification of breast lesions in ultrasound images. Phys Med Biol 62(19):7714
Hijab A, Rushdi MA, Gomaa MM, Eldeib A (2019) Breast cancer classification in ultrasound images using transfer learning. Fifth International Conference on Advances in Biomedical Engineering (ICABME) 1–4. IEEE
Hu Y, Guo Y, Wang Y, Yu J, Li J, Zhou S, Chang C (2019) Automatic tumor segmentation in breast ultrasound images using a dilated fully convolutional network combined with an active contour model. Med Phys 46(1):215–228
Klimonda Z, Karwat P, Dobruch-Sobczak K, Piotrzkowska-Wróblewska H, Litniewski J (2019) Breast-lesions characterization using quantitative ultrasound features of peritumoral tissue. Sci Rep 9(1):1–9
Koundal D, Gupta S, Singh S (2016) Speckle reduction method for thyroid ultrasound images in neutrosophic domain. IET Image Proc 10(2):167–175
Lankton S, Tannenbaum A (2008) Localizing region-based active contours. IEEE Trans Image Process 17(11):2029–2039
Liu S, Wang Y, Yang X, Lei B, Liu L, Li SX, Ni D, Wang T (2019) Deep learning in medical ultrasound analysis: a review. Engineering 5(2):261–275
Masud M, Eldin Rashed AE, Hossain MS (2020) Convolutional neural network-based models for diagnosis of breast cancer. Neural Comput Appl 9:1–2
Moon WK, Lee YW, Ke HH, Lee SH, Huang CS, Chang RF (2020) Computer-aided diagnosis of breast ultrasound images using ensemble learning from convolutional neural networks. Comput Methods Programs Biomed 190:105361
Muduli D, Dash R, Majhi B (2022) Automated diagnosis of breast cancer using multi-modal datasets: a deep convolution neural network based approach. Biomed Signal Process Control 71:102825
Ronneberger O, Fischer P, Brox T (2015) U-Net: convolutional networks for biomedical image segmentation. International conference on medical image computing and computer-assisted intervention. Springer, Cham, pp 234–241
Saba T, Abunadi I, Sadad T, Khan AR, Bahaj SA (2022) Optimizing the transfer-learning with pretrained deep convolutional neural networks for first stage breast tumor diagnosis using breast ultrasound visual images. Microsc Res Tech 85(4):1444–1453
Salama AA, Smarandache F, Eisa M (2014) Introduction to image processing via neutrosophic techniques. Infinite study
Salama AA, Smarandache F, ElGhawalby H (2018) Neutrosophic approach to grayscale images domain. Neutrosophic Sets Syst 21:13–19
Sezer A, Sezer HB (2020) Deep convolutional neural network-based automatic classification of neonatal hip ultrasound images: a novel data augmentation approach with speckle noise reduction. Ultrasound Med Biol 46(3):735–749
Sivanandan R, Jayakumari J (2020) An improved ultrasound tumor segmentation using CNN activation map clustering and active contours. IEEE 5th International Conference on Computing Communication and Automation (ICCCA). IEEE, pp 263–268
Sivanandan R, Jayakumari J (2020) Neutrosophic texture-region difference-based fuzzy c-means clustering of ultrasound tumor images. Biomed Eng: Appl Basis Commun 32(06):2050049
Tanaka H, Chiu SW, Watanabe T, Kaoku S, Yamaguchi T (2019) Computer-aided diagnosis system for breast ultrasound images using deep learning. Phys Med Biol 64(23):235013
Torrey L, Shavlik J (2010) Transfer learning. Handbook of research on machine learning applications and trends: algorithms, methods, and techniques pp. IGI global, pp 242–264
Vakanski A, Xian M, Freer PE (2020) Attention-enriched deep learning model for breast tumor segmentation in ultrasound images. Ultrasound Med Biol 46(10):2819–2833
Wu H, Zhang J, Huang K, Liang K, Yu Y (2019) FastFCN: rethinking dilated convolution in the backbone for semantic segmentation. arXiv preprint arXiv:1903.11816
Yoon HJ, Gounley J, Gao S, Alawad M, Ramanathan A, Tourassi G (2019) Model-based hyperparameter optimization of convolutional neural networks for information extraction from cancer pathology reports on HPC. IEEE EMBS International Conference on Biomedical & Health Informatics (BHI), pp 1–4
Zhang M, Li H, Lyu J, Ling SH, Su S (2019) Multi-level CNN for lung nodule classification with gaussian process assisted hyperparameter optimization. arXiv preprint arXiv:1901.00276
Zhang Z, Li Y, Wu W, Chen H, Cheng L, Wang S (2019) Tumor detection using deep learning method in automated breast ultrasound. Biomed Signal Process Control 68:102677
Zhuang Z, Li N, Joseph Raj AN, Mahesh VG, Qiu S (2019) An RDAU-NET model for lesion segmentation in breast ultrasound images. PLoS ONE 14(8):e0221535
Zuluaga-Gomez J, Al Masry Z, Benaggoune K, Meraghni S, Zerhouni N (2021) A CNN-based methodology for breast cancer diagnosis using thermal images. Comput Methods Biomech Biomed Eng Imaging Vis 9(2):131–145
Author information
Authors and Affiliations
Contributions
Revathy Sivanandan: Methodology, Software, Writing-Original draft preparation, Visualization, Investigation, Validation.
Jayakumari J: Supervision, Validation, Writing- Reviewing and Editing.
Corresponding author
Ethics declarations
Conflicts of interest/Competing interests
None.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Sivanandan, R., J, J. Bayesian optimized novel CNN for improved diagnosis from ultrasound breast tumor images. Multimed Tools Appl 82, 22815–22833 (2023). https://doi.org/10.1007/s11042-023-14468-0
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11042-023-14468-0