Classification of Benign and Malignant Masses in Ultrasound Breast Image Based on Geometric and Echo Features

Fukuoka, Daisuke; Ikedo, Yuji; Hara, Takeshi; Fujita, Hiroshi; Takada, Etsuo; Endo, Tokiko; Morita, Takako

doi:10.1007/978-3-540-70538-3_60

Cited by 7 publications

(7 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They were compactness, orientation, shape, depth-to-width ratio, homogeneity and posterior analysis. These six features were similar to the features selected in our previous study [3] and were based on the image features that radiologists' found useful and routinely consulted in breast ultrasound images interpretations. A summary of the six features is given in the next paragraph.…”

Section: 5-dimensional Analysismentioning

confidence: 94%

See 1 more Smart Citation

Classifying Breast Masses in Volumetric Whole Breast Ultrasound Data: A 2.5-Dimensional Approach

Okada

Fukuoka

Muramatsu

et al. 2010

Digital Mammography

Self Cite

View full text Add to dashboard Cite

Section: 5-dimensional Analysismentioning

confidence: 94%

“…The volumetric ultrasound data of a whole breast consist of a stack of two-dimensional images, each depicting an axial slice image of the breast. In exploiting the benefit of volumetric data, three-dimensional analysis was used in our previous study in classifying malignant and benign breast masses [3].…”

Section: Introductionmentioning

confidence: 99%

Classifying Breast Masses in Volumetric Whole Breast Ultrasound Data: A 2.5-Dimensional Approach

Okada

Fukuoka

Muramatsu

et al. 2010

Digital Mammography

Self Cite

View full text Add to dashboard Cite

“…Sonograph in breast cancer diagnosis field has been used for differentiating solid from cyst masses. The role of ultrasound image has been expanded by improving the quality of the image and now ultrasounds have become a complementary test to mammographs for differentiating benign from malignant masses (Lee et al, 2008;Song et al, 2005). Unlike mammography, ultrasound can deal with dense breast tissue.…”

Section: Introductionmentioning

confidence: 99%

“…Unlike mammography, ultrasound can deal with dense breast tissue. Therefore, it is highly recommended for women with dense breast tissue or women under 40 years of age to take an ultrasound examination (Lee et al, 2008). Ultrasound image is processed carefully to differentiate malignant from benign masses.…”

Section: Introductionmentioning

confidence: 99%

Ultrasound based computer aided diagnosis of breast cancer: Evaluation of a new feature of mass central regularity degree

Al-Yousef

Samarasinghe²

2011

Chan, F., Marinova, D. And Anderssen, R.S. (Eds) MODSIM2011, 19th International Congress on Modelling and Simulation.

View full text Add to dashboard Cite

Breast cancer is the second leading cause of death after lung cancer in women all over the world. The survival rate of breast cancer patients depends on the stage of diagnosis; patients with stage 0 are more likely to reach cancer free state. Therefore, early detection of breast cancer is the key to patient survival. In order to enhance diagnostic accuracy of breast cancer, computer aided diagnosis (CAD) systems have been built. Ultrasound is one of the most frequently used methods for early detection of breast cancer. Currently, the accuracy of CAD systems based on ultrasound images is about 90% and needs further enhancement in order to save lives of the undetected. A meaningful approach to do this is to explore new and meaningful features with discriminating ability and incorporate them into CAD systems. Recently, from a thorough investigation of the images, we extracted a new geometric feature related to the mass shape in ultrasound images called Central Regularity Degree (CRD). The CRD reflects the degree of regularity of the middle part of the mass. To demonstrate the effect of CRD on differentiating malignant from benign masses and the potential improvement to the diagnostic accuracy of breast cancer using ultrasound, this study evaluated the diagnostic accuracy of different classifiers when the CRD was added to five powerful mass features obtained from previous studies including one geometric feature: Depth-Width ratio (DW); two morphological features: shape and margin; blood flow and age. Artificial Neural Networks (ANN), K Nearest Neighbour (KNN), Nearest Centroid, Linear Discriminant Analysis (LDA), and Receiver Operating Characteristic (ROC) analysis were employed for classification and evaluation. Ninety nine breast sonograms-46 malignant and 53 benign-were evaluated. The results reveal that CRD is an effective feature discriminating between malignant and benign cases leading to improved accuracy of diagnosis of breast cancer. The best results were obtained by ANN where the area under ROC curve (Az) for training and testing using all features except CRD was 100% and 81.8%, respectively, and 100% and 95.45% using all features. Therefore, the overall improvement by adding CRD was about 14%, a significant improvement.

show abstract

“…Compared to manual methods, semi-automatic methods only require sparse seeds, such as user clicks, for localization. For example, the method presented in [10] assumes manually segmented masses and uses those as input for subsequent processes. Most of these conventional approaches focus on developing effective features in distinguishing benign and malignant masses.…”

mentioning

confidence: 99%

Joint Weakly and Semi-Supervised Deep Learning for Localization and Classification of Masses in Breast Ultrasound Images

Shin

Lee

Yun

et al. 2019

IEEE Trans. Med. Imaging

127

View full text Add to dashboard Cite

We propose a framework for localization and classification of masses in breast ultrasound (BUS) images. We have experimentally found that training convolutional neural network based mass detectors with large, weakly annotated datasets presents a non-trivial problem, while overfitting may occur with those trained with small, strongly annotated datasets. To overcome these problems, we use a weakly annotated dataset together with a smaller strongly annotated dataset in a hybrid manner. We propose a systematic weakly and semi-supervised training scenario with appropriate training loss selection. Experimental results show that the proposed method can successfully localize and classify masses with less annotation effort. The results trained with only 10 strongly annotated images along with weakly annotated images were comparable to results trained from 800 strongly annotated images, with the 95% confidence interval of difference -3.00%-5.00%, in terms of the correct localization (CorLoc) measure, which is the ratio of images with intersection over union with ground truth higher than 0.5. With the same number of strongly annotated images, additional weakly annotated images can be incorporated to give a 4.5% point increase in CorLoc, from 80.00% to 84.50% (with 95% confidence intervals 76.00%-83.75% and 81.00%-88.00%). The effects of different algorithmic details and varied amount of data are presented through ablative analysis.Index Terms-Breast ultrasound, convolutional neural networks, mass classification, mass localization, semisupervised learning, weakly supervised learning.

show abstract

Classification of Benign and Malignant Masses in Ultrasound Breast Image Based on Geometric and Echo Features

Cited by 7 publications

References 10 publications

Classifying Breast Masses in Volumetric Whole Breast Ultrasound Data: A 2.5-Dimensional Approach

Classifying Breast Masses in Volumetric Whole Breast Ultrasound Data: A 2.5-Dimensional Approach

Ultrasound based computer aided diagnosis of breast cancer: Evaluation of a new feature of mass central regularity degree

Joint Weakly and Semi-Supervised Deep Learning for Localization and Classification of Masses in Breast Ultrasound Images

Contact Info

Product

Resources

About