Dong Yifeng
ABSTRACT
Breast cancer is one of the cancers with highest mortality rate, which are impairing the health of millions of women globally. Nowadays, the detection method of breast tumors based on breast ultrasound images (BUI) is semi-automatic. The localization of a region of interest (ROI) and the delineation of breast tumors within ROI need to be manually performed by an experienced doctor. In this work, we propose a fully automatic approach based on semantic segmentation to add contextual information in the complete image interpretation and avoid the ROI definition. The densely connected convolutional neural network is used for this approach to generate the feature maps. Dataset with 200 breast tumor images from Institute of Oncology in Warsaw was used to evaluate the performance of this approach. For DenseNet, the accuracy of our method is 99.2%. For traditional fully convolutional network, this accuracy is 85.60%. For the results of segmentation, the dice coefficient is equal to 0.8329. The results have demonstrated that the proposed method is accurate and objective for the segmentation of breast tumors in BUIs. Moreover, the robustness and generalization capacity of the method have been proven in this dataset.
- Siegel R L, Miller K D, Jemal A, Cancer statistics[J]. CA: A Cancer Journal for Clinicians, 2015, 65: 5-29.Google Scholar
- Fan L, Strasser W K, Li J J, Louis J S, Finkelstein D M, Breast in China[J], Lancet Oncol. 2014,15(7): e279-289.Google ScholarCross Ref
- Drukker K, Giger M L, Vyborny C J, Computerized detection and classification of cancer on breast ultrasound[J]. Academic Radiology, 2004, 11(5): 526-535.Google ScholarCross Ref
- Shi C, Cheng Y, Liu F, Wang Y, Bai J, Tamura S, A hierarchical local region-based sparse shape composition for liver segmentation in CT scans[J], Pattern Recognition, 2016, 50: 88-106.Google ScholarDigital Library
- Shi C, Cheng Y, Wang J, Low-rank and sparse decomposition based shape model and probabilistic atlas for automatic pathological organ segmentation[J]. Medical Image Analysis, 2017, 38(7): 30.Google ScholarCross Ref
- Cheng H D, Shan J, Ju W, Automated breast cancer detection and classification using ultrasound images: a survey[J]. Pattern Recognition, 2010, 43(1): 299-317.Google ScholarDigital Library
- Xian M, Zhang Y, Cheng H D, Breast ultrasound image segmentation: a survey[J]. International Journal of Computer Assisted Radiology & Surgery, 2018, 12(3): 1-15.Google Scholar
- Yap M H, Edirisinghe E A, Bez H E, Fully automatic lesion boundary detection in ultrasound breast images[C]. The International Society for Optical Engineering, 2007, 6512:10.1117/12.708625.Google Scholar
- Yap M H, Edirisinghe E A, Bez H E, A novel algorithm for initial lesion detection in ultrasound breast images[J]. Journal of Applied Clinical Medical Physics, 2008, 9(4): 2741.Google ScholarCross Ref
- Kwak J, Kim S, Kim N, Rd-based seeded region growing for extraction of breast tumor in an ultrasound volume[J]. Springer, 2005, pp: 799-808.Google ScholarDigital Library
- Huang Y L, Chen D R, Watershed segmentation for breast tumor in 2D sonography[J], Ultrasound in Medicine & Biology, 2004, 30(5): 625-632.Google ScholarCross Ref
- Zhou Z, Wu W, Wu S, Semi-automatic breast ultrasound image segmentation based on mean shift and graph cuts[J], Ultrasonic Imaging, 2014, 36(4): 256.Google ScholarCross Ref
- Srinivasan S S, Swaminathan R, Semi automatic segmentation of breast thermograms using variational level set method[C]. The 15st International Conference on Biomedical Engineering: Springer, 2014, pp: 231–234.Google Scholar
- Boukerroui D, Basset O, Nicole G, Multiresolution texture based adaptive clustering algorithm for breast lesion segmentation[J]. Ultrasound in Medicine & Biology, 1998, 8(2): 135-144.Google Scholar
- Boukerroui D, Baskurt A, Noble J A, Segmentation of ultrasound image multiresolution 2D and 3D algorithm based on global and local statistics[J]. Pattern Recognition Letters, 2003, 24(4/5): 779-790.Google ScholarDigital Library
- Gou X, Brady M, Segmentation of ultrasound B-mode images with intensity inhomogeneity correction.[J]. IEEE Transactions on Medical Imaging, 2002, 21(1): 48-48.Google ScholarCross Ref
- Yezzi A, Kichenassamy S, Kumar A, A geometric snake model for segmentation of medical imagery[J]. IEEE Transactions on Medical Imaging, 2002, 16(2): 199-209.Google ScholarCross Ref
- Ashton E A, Parker K J, Multiple resolution bayesian segmentation of ultrasound images[J]. Ultrasonic Imaging, 1995, 17(4): 291-304.Google ScholarCross Ref
- Xian M, Zhang Y, Cheng H D, Fully automatic segmentation of breast ultrasound images based on breast characteristics in space and frequency domains[J]. Pattern Recognition, 2015, 48(2): 485-497.Google ScholarDigital Library
- Madabhushi A, Metaxas D N, Automatic boundary extraction of ultrasonic breast lesions[C]. IEEE International Symposium on Biomedical Imaging: IEEE, 2002, pp: 601-604.Google ScholarCross Ref
- Ikedo Y, Fukuoka D, Hara T, Development of a fully automatic scheme for detection of masses in whole breast ultrasound images[J]. Medical Physics, 2007, 34(11): 4378-4388.Google ScholarCross Ref
- Huang Y L, Chen D R, Watershed segmentation for breast tumor in 2D sonography[J]. Ultrasound in Medicine & Biology, 2004, 30(5): 625-632.Google ScholarCross Ref
- Almajalid R, Shan J, Du Y, Development of a deep-learning-based method for breast ultrasound image segmentation[C]. 2018 17st IEEE International Conference on Machine Learning and Applications: IEEE, 2018, pp: 1103-1108.Google ScholarCross Ref
- Zhang Y, Liu Y, Cheng H, Fully multi-target segmentation for breast ultrasound image based on fully convolutional network[J]. Medical & Biological Engineering & Computing, 2020, 58(1): 2049–2061.Google ScholarCross Ref
- Huang G, Liu Z, Laurens V D M, Densely connected convolutional networks[C]. IEEE Conference on Computer Vision and Pattern Recognition:IEEE, 2016, pp: 2261-2269.Google Scholar
- Pons G, Joan M, Robert M, Breast-lesion segmentation combining B-mode and elastography ultrasound[J]. Ultrasonic Imaging, 2015, 38(2016): 209-224.Google Scholar
- Xian M, Huang J, Zhang Y, Multiple-domain knowledge based MRF model for tumor segmentation in breast ultrasound images[C]. International Conference on Image Processing:IEEE, 2012, pp: 2021-2024.Google Scholar
Recommendations
Neoadjuvant chemotherapy response evaluation in breast cancer based on mammogram registration and tumor segmentation
The standard approach for assessing the response of breast tumors to neoadjuvant chemotherapy is to monitor gross changes in tumor size as measured by physical exam and/or conventional imaging such as mammography. However, the deformable nature of the ...
RETRACTED ARTICLE: Classification of ultrasound breast cancer tumor images using neural learning and predicting the tumor growth rate
AbstractUltrasound breast cancer tumor growth model is dependence on the cancer tumor growth size on time. Breast cancer tumor progresses on its growth and evaluated to estimate the survival time. The volume of the tumor changes because of the cell growth ...
Breast tumor detection and classification based on density
AbstractBreast cancer is the most widely disease in women and is considered one of the biggest causes of death in women. Early detection, classification, and diagnosis of this cancer are essential to reduce the death cases. The efficiency of the diagnosis ...
Comments