A non-linear morphometric feature selection approach for breast tumor contour from ultrasonic images

Pereira, Wagner Coelho de Albuquerque; Alvarenga, André Victor; Infantosi, Antonio Fernando Catelli; Macrini, Leonardo; Pedreira, Carlos Eduardo

doi:10.1016/j.compbiomed.2010.10.003

Cited by 22 publications

(11 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Among different proposed approaches considering solid mass classification, there are two main feature descriptors, 27 i.e., echo texture 11,28 and shape and margin features. 29 We present a couple of works on multistage machine learning methods. For a full review, please refer to Cheng et al 26 Liu et al 30 proposed a breast classification system for color Doppler flow imaging and B-mode ultrasound.…”

Section: Breast Ultrasound Classification Approachesmentioning

confidence: 99%

Breast ultrasound lesions recognition: end-to-end deep learning approaches

et al. 2018

View full text Add to dashboard Cite

Multistage processing of automated breast ultrasound lesions recognition is dependent on the performance of prior stages. To improve the current state of the art, we propose the use of end-to-end deep learning approaches using fully convolutional networks (FCNs), namely FCN-AlexNet, FCN-32s, FCN-16s, and FCN-8s for semantic segmentation of breast lesions. We use pretrained models based on ImageNet and transfer learning to overcome the issue of data deficiency. We evaluate our results on two datasets, which consist of a total of 113 malignant and 356 benign lesions. To assess the performance, we conduct fivefold cross validation using the following split: 70% for training data, 10% for validation data, and 20% testing data. The results showed that our proposed method performed better on benign lesions, with a top "mean Dice" score of 0.7626 with FCN-16s, when compared with the malignant lesions with a top mean Dice score of 0.5484 with FCN-8s. When considering the number of images with Dice score >0.5, 89.6% of the benign lesions were successfully segmented and correctly recognised, whereas 60.6% of the malignant lesions were successfully segmented and correctly recognized. We conclude the paper by addressing the future challenges of the work.

show abstract

Section: Breast Ultrasound Classification Approachesmentioning

confidence: 99%

Breast ultrasound lesions recognition: end-to-end deep learning approaches

et al. 2018

View full text Add to dashboard Cite

show abstract

“…Nevertheless, the interobserver agreement kappa indices described in the literature range from 0.28 to 0.83 for a diagnosis essentially based on the real-time subjective evaluation of the morphological aspects of a lesion (10)(11)(12) . Aiming to reduce the number of biopsies taken from benign solid tumors and increase the consistency of the diagnostic interpretation of ultrasound, several studies have proposed methods to aid in the diagnosis of breast cancer (15,(20)(21)(22)(23)(24) .…”

Section: Discussionmentioning

confidence: 99%

“…However, because of the high demand for breast ultrasound, ultrasound examinations of the breast are often performed by general radiologists or ultrasonographers with limited experience in breast imaging, rather than breast specialists, which reduces the sensitivity of the method and increases the number of false-positive results ( 9 - 13 ) . The growing concern of doctors and patients about the increased risk of breast cancer related to high breast density, combined with the limitations of mammography in dense breasts, has led to the development of additional screening tools, automated breast ultrasound (ABUS) being one such tool ( 13 - 15 ) .…”

Section: Introductionmentioning

confidence: 99%

Preliminary study of the technical limitations of automated breast ultrasound: from procedure to diagnosis

et al. 2020

View full text Add to dashboard Cite

Objective: To evaluate the main technical limitations of automated breast ultrasound and to determine the proportion of examinations excluded. Materials and Methods: We evaluated 440 automated breast ultrasound examinations performed, over a 12-month period, by technicians using an established protocol. Results: In five cases (1.1%), the examination was deemed unacceptable for diagnostic purposes, those examinations therefore being excluded. Conclusion: Automated breast ultrasound is expected to overcome some of the major limitations of conventional ultrasound in breast cancer screening. In Brazil, this new method can be accepted for inclusion in routine clinical practice only after its advantages have been validated in the national context.

show abstract

“…The infiltrative nature of malignant tumors generate an irregular pattern of impedance discontinuities, which results irregular, spiculated or ill-defined boundary in breast ultrasound images. However, benign tumors have a more uniform growth with smooth, round, and well-defined boundaries [9]. Based on this, Wu et al [17,18] combined textural and morphological features and achieved an A Z value of 0.9614.…”

Section: Discussionmentioning

confidence: 99%

“…It is essential to quantify the characteristics of breast tumors for the detection as well as discrimination, which are quite often difficult to grasp due to intrinsic limitations of the ultrasound imaging process, such as low contrast, speckle noise, heterogeneity or artifacts. It is significant to explore the feature, or set of features, which provide better quantifications of the characteristics of tumors [9]. Some general guidelines [10] for identifying significant features which leads to accurate diagnosis are discrimination, reliability, independence and optimality.…”

Section: Introductionmentioning

confidence: 99%

Assessment of combined textural and morphological features for diagnosis of breast masses in ultrasound

Prabusankarlal

Thirumoorthy²,

Manavalan³

2015

Hum. Cent. Comput. Inf. Sci.

View full text Add to dashboard Cite

The objective of this study is to assess the combined performance of textural and morphological features for the detection and diagnosis of breast masses in ultrasound images. We have extracted a total of forty four features using textural and morphological techniques. Support vector machine (SVM) classifier is used to discriminate the tumors into benign or malignant. The performance of individual as well as combined features are assessed using accuracy(Ac), sensitivity(Se), specificity(Sp), Matthews correlation coefficient(MCC) and area A Z under receiver operating characteristics curve. The individual features produced classification accuracy in the range of 61.66% and 90.83% and when features from each category are combined, the accuracy is improved in the range of 79.16% and 95.83%. Moreover, the combination of gray level co-occurrence matrix (GLCM) and ratio of perimeters (P ratio) presented highest performance among all feature combinations (Ac 95.85%, Se 96%, Sp 91.46%, MCC 0.9146 and A Z 0.9444).The results indicated that the discrimination performance of a computer aided breast cancer diagnosis system increases when textural and morphological features are combined.

show abstract

A non-linear morphometric feature selection approach for breast tumor contour from ultrasonic images

Cited by 22 publications

References 26 publications

Breast ultrasound lesions recognition: end-to-end deep learning approaches

Breast ultrasound lesions recognition: end-to-end deep learning approaches

Preliminary study of the technical limitations of automated breast ultrasound: from procedure to diagnosis

Assessment of combined textural and morphological features for diagnosis of breast masses in ultrasound

Contact Info

Product

Resources

About