• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.134-136
• /
• 2021

Research related to early diagnosis of breast cancer using artificial intelligence algorithms has been actively conducted in recent years. Although various algorithms that classify breast cancer based on a few publicly available ultrasound breast cancer images have been published, these methods show various limitations such as, processing speed and accuracy suitable for the user's purpose. To solve this problem, in this paper, we propose a multi-stage transfer learning where ResNet model trained on ImageNet is transfer learned to microscopic cancer cell line images, which was again transfer learned to classify ultrasound breast cancer images as benign and malignant. The images for the experiment consisted of 250 breast cancer ultrasound images including benign and malignant images and 27,200 cancer cell line images. The proposed multi-stage transfer learning algorithm showed more than 96% accuracy when classifying ultrasound breast cancer images, and is expected to show higher utilization and accuracy through the addition of more cancer cell lines and real-time image processing in the future.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.3
• /
• pp.678-700
• /
• 2023

Diagnosing breast diseases using ultrasound (US) images remains challenging because it is time-consuming and requires expert radiologist knowledge. As a result, the diagnostic performance is significantly biased. To assist radiologists in this process, computer-aided diagnosis (CAD) systems have been developed and used in practice. This type of system is used not only to assist radiologists in examining breast ultrasound images (BUS) but also to ensure the effectiveness of the diagnostic process. In this study, we propose a new approach for breast lesion localization and segmentation using a multi-scale pyramid of the ultrasound image of a breast organ and a convolutional semantic segmentation network. Unlike previous studies that used only a deep detection/segmentation neural network on a single breast ultrasound image, we propose to use multiple images generated from an input image at different scales for the localization and segmentation process. By combining the localization/segmentation results obtained from the input image at different scales, the system performance was enhanced compared with that of the previous studies. The experimental results with two public datasets confirmed the effectiveness of the proposed approach by producing superior localization/segmentation results compared with those obtained in previous studies.

• Journal of information and communication convergence engineering
• /
• v.20 no.1
• /
• pp.58-64
• /
• 2022

Breast ultrasonic reading is critical as a primary screening test for the early diagnosis of breast cancer. However, breast ultrasound examinations show significant differences in diagnosis based on the difference in image quality according to the ultrasonic equipment, experience, and proficiency of the examiner. Accordingly, studies are being actively conducted to analyze the texture characteristics of normal breast tissue, positive tumors, and malignant tumors using breast ultrasonography and to use them for computer-assisted diagnosis. In this study, breast ultrasonography was conducted to select 247 ultrasound images of 71 normal breast tissues, 87 fibroadenomas among benign tumors, and 89 malignant tumors. The selected images were calculated using a statistical method with 21 feature parameters extracted using the gray level co-occurrence matrix algorithm, and classified as normal breast tissue, benign tumor, and malignancy. In addition, we proposed five feature parameters that are available for computer-aided diagnosis of breast cancer classification. The average classification rate for normal breast tissue, benign tumors, and malignant tumors, using this feature parameter, was 82.8%.

• International Journal of Internet, Broadcasting and Communication
• /
• v.16 no.1
• /
• pp.124-131
• /
• 2024

Breast cancer classification using ultrasound, while widely employed, faces challenges due to its relatively low predictive value arising from significant overlap in characteristics between benign and malignant lesions, as well as operator-dependency. To alleviate these challenges and reduce dependency on radiologist interpretation, the implementation of automatic breast cancer classification in ultrasound image can be helpful. To deal with this problem, we propose a semi-supervised deep learning framework for breast cancer classification. In the proposed method, we could achieve reasonable performance utilizing less than 50% of the training data for supervised learning in comparison to when we utilized a 100% labeled dataset for training. Though it requires more modification, this methodology may be able to alleviate the time-consuming annotation burden on radiologists by reducing the number of annotation, contributing to a more efficient and effective breast cancer detection process in ultrasound images.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.10a
• /
• pp.655-657
• /
• 2021

It is challenging to find breast ultrasound image training dataset to develop an accurate machine learning model due to various regulations, personal information issues, and expensiveness of acquiring the images. However, studies targeting transfer learning for ultrasound breast cancer images classification have not been able to achieve high performance compared to radiologists. Here, we propose an improved transfer learning model for ultrasound breast cancer classification using publicly available dataset. We argue that with a proper combination of ImageNet pre-trained model and optimizer, a better performing model for ultrasound breast cancer image classification can be achieved. The proposed model provided a preliminary test accuracy of 99.5%. With more experiments involving various hyperparameters, the model is expected to achieve higher performance when subjected to new instances.

• International journal of advanced smart convergence
• /
• v.8 no.1
• /
• pp.24-34
• /
• 2019

In this research, a practical deep learning framework to differentiate the lesions and nodules in breast acquired with ultrasound imaging has been proposed. 7408 ultrasound breast images of 5151 patient cases were collected. All cases were biopsy proven and lesions were semi-automatically segmented. To compensate for the shift caused in the segmentation, the boundaries of each lesion were drawn using Fully Convolutional Networks(FCN) segmentation method based on the radiologist's specified point. The data set consists of 4254 benign and 3154 malignant lesions. In 7408 ultrasound breast images, the number of training images is 6579, and the number of test images is 829. The margin between the boundary of each lesion and the boundary of the image itself varied for training image augmentation. The training images were augmented by varying the margin between the boundary of each lesion and the boundary of the image itself. The images were processed through histogram equalization, image cropping, and margin augmentation. The networks trained on the data with augmentation and the data without augmentation all had AUC over 0.95. The network exhibited about 90% accuracy, 0.86 sensitivity and 0.95 specificity. Although the proposed framework still requires to point to the location of the target ROI with the help of radiologists, the result of the suggested framework showed promising results. It supports human radiologist to give successful performance and helps to create a fluent diagnostic workflow that meets the fundamental purpose of CADx.

• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.24
• /
• pp.10739-10743
• /
• 2015

The present study aimed at evaluating and comparing the diagnostic performance of B-mode ultrasound (US), elastography score (ES), and strain ratio (SR) for the differentiation of breast lesions. This retrospective study enrolled 431 lesions from 417 in-hospital patients. All patients were examined with both conventional ultrasound and elastography. Two experienced radiologists reviewed ultrasound and elasticity images. The histopathologic result obtained from ultrasound-guided core biopsy or operation excisions were used as the reference standard. Pathologic examination revealed 276 malignant lesions (64%) and 155 benign lesions (36%). A cut-off point of 4.15 (area under the curve, 0.891) allowed significant differentiation of malignant and benign lesions. ROC (receiver-operating characteristic) curves showed a higher value for combination of B-mode ultrasound and elastography for the diagnosis of breast lesions. Conventional ultrasound combined elastography showed high sensitivity, specificity, and accuracy for group II lesions (10mm${\leq}20mm$). Elastography combined with conventional ultrasound show high specificity and accuracy for differentiation of benign and malignant breast lesions. Elastography is particularly important for the diagnosis of BI-RADS 4 and small breast lesions.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.66 no.8
• /
• pp.1272-1277
• /
• 2017

To assist radiologists for the characterization of breast masses, Computer-aided Diagnosis(CADx) system has been studied. The CADx system can improve the diagnostic accuracy of radiologists by providing objective information about breast masses. Morphological and texture features were extracted from the breast ultrasound images. Based on extracted features, the CADx system retrieves masses that are similar to a query mass from a reference library using a k-nearest neighbor (k-NN) approach. Eight similarity measures of distance, Euclidean, Chebyshev(Minkowski family), Canberra, Lorentzian($F_2$ family), Wave Hedges, Motyka(Intersection family), and Cosine, Dice(Inner Product family) are evaluated by ROC(Receiver Operating Characteristic) analysis. The Inner Product family measure used with the k-NN classifier provided slightly higher performance for classification of malignant and benign masses than those with the Minkowski, $F_2$, and Intersection family measures.

• Korean Journal of Radiology
• /
• v.25 no.2
• /
• pp.134-145
• /
• 2024

Abnormalities on breast ultrasound (US) images which do not meet the criteria for masses are referred to as nonmass lesions. These features and outcomes have been investigated in several studies conducted by Asian researchers. However, the term "nonmass" is not included in the American College of Radiology (ACR) Breast Imaging Reporting and Data System (BI-RADS) 5th edition for US. According to the Japan Association of Breast and Thyroid Sonology guidelines, breast lesions are divided into mass and nonmass. US findings of nonmass abnormalities are classified into five subtypes: abnormalities of the ducts, hypoechoic areas in the mammary glands, architectural distortion, multiple small cysts, and echogenic foci without a hypoechoic area. These findings can be benign or malignant; however, focal or segmental distributions and presence of calcifications suggest malignancy. Intraductal, invasive ductal, and lobular carcinomas can present as nonmass abnormalities. For the nonmass concept to be included in the next BI-RADS and be widely accepted in clinical practice, standardized terminologies, an interpretation algorithm, and outcome-based evidence are required for both screening and diagnostic US.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.10a
• /
• pp.307-310
• /
• 2021

In the recent years, studies using Computer-Aided Diagnostics(CAD) have been actively conducted, such as signal and image processing technology using breast ultrasound images, automatic image optimization technology, and automatic detection and classification of breast masses. As computer diagnostic technology is developed, it is expected that early detection of cancer will proceed accurately and quickly, reducing health insurance and test ice for patients, and eliminating anxiety about biopsy. In this paper, a quantitative analysis of tumors was conducted in ultrasound images using a gray level co-occurrence matrix(GLCM) to experiment with the possibility of use for computer assistance diagnosis.