Fully-automated deep learning pipeline for segmentation and classification of breast ultrasound images

Podda, Alessandro Sebastian; Balia, Riccardo; Barra, Silvio; Carta, Salvatore; Fenu, Gianni; Piano, Leonardo

doi:10.1016/j.jocs.2022.101816

Cited by 27 publications

(4 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Dataset 1 (BUSI), the most widely used publicly available dataset, contains several anomalous cases confirmed by two experienced radiologists. Podda et al (2022) also reported such outliers in their work. We have identified 84 anomalous cases among the 647 benign and malignant BUS images in dataset 1 that affect the classification performance significantly (see table 3 in section 4).…”

Section: Dataset Outliersmentioning

confidence: 81%

CAM-QUS guided self-tuning modular CNNs with multi-loss functions for fully automated breast lesion classification in ultrasound images

Tasnim,

Hasan

2023

Phys. Med. Biol.

View full text Add to dashboard Cite

Objective. Breast cancer is the major cause of cancer death among women worldwide. Deep learning-based computer-aided diagnosis (CAD) systems for classifying lesions in breast ultrasound images can help materialise the early detection of breast cancer and enhance survival chances. Approach. This paper presents a completely automated BUS diagnosis system with modular convolutional neural networks tuned with novel loss functions. The proposed network comprises a dynamic channel input enhancement network, an attention-guided InceptionV3-based feature extraction network, a classification network, and a parallel feature transformation network to map deep features into quantitative ultrasound (QUS) feature space. These networks function together to improve classification accuracy by increasing the separation of benign and malignant class-specific features and enriching them simultaneously. Unlike the categorical crossentropy (CCE) loss-based traditional approaches, our method uses two additional novel losses: class activation mapping (CAM)-based and QUS feature-based losses, to capacitate the overall network learn the extraction of clinically valued lesion shape and texture-related properties focusing primarily the lesion area for explainable AI (XAI). Main results. Experiments on four public, one private, and a combined breast ultrasound dataset are used to validate our strategy. The suggested technique obtains an accuracy of 97.28%, sensitivity of 93.87%, F1-score of 95.42% on dataset 1 (BUSI), and an accuracy of 91.50%, sensitivity of 89.38%, and F1-score of 89.31% on the combined dataset, consisting of 1494 images collected from hospitals in five demographic locations using four ultrasound systems of different manufacturers. These results outperform techniques reported in the literature by a considerable margin. Significance. The proposed CAD system provides diagnosis from the auto-focused lesion area of B-mode BUS images, avoiding the explicit requirement of any segmentation or region of interest extraction, and thus can be a handy tool for making accurate and reliable diagnoses even in unspecialized healthcare centers.

show abstract

Section: Dataset Outliersmentioning

confidence: 81%

CAM-QUS guided self-tuning modular CNNs with multi-loss functions for fully automated breast lesion classification in ultrasound images

Tasnim,

Hasan

2023

Phys. Med. Biol.

View full text Add to dashboard Cite

show abstract

“…In TL, the weight-sharing mechanism is achieved by freezing the weights of some layers in the existing network and then retraining for newly introduced layers. This procedure is successfully being adopted to solve several problems, such as human liver cancer drug response, HBV, skin cancer diagnosis, and plant disease predictions [11,[22][23][24][25].…”

Section: The Resnet101 Transfer Learningmentioning

confidence: 99%

Developing a Breast Cancer Resistance Protein Substrate Prediction System Using Deep Features and LDA

Hassan,

Ali,

Young Kim

et al. 2023

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

Breast cancer resistance protein (BCRP) is an important resistance protein that significantly impacts anticancer drug discovery, treatment, and rehabilitation. Early identification of BCRP substrates is quite a challenging task. This study aims to predict early substrate structure, which can help to optimize anticancer drug development and clinical diagnosis. For this study, a novel intelligent approach-based methodology is developed by modifying the ResNet101 model using transfer learning (TL) for automatic deep feature (DF) extraction followed by classification with linear discriminant analysis algorithm (TLRNDF-LDA). This study utilized structural fingerprints, which are exploited by DF contrary to conventional molecular descriptors. The proposed in silico model achieved an outstanding accuracy performance of 98.56% on test data compared to other state-of-the-art approaches using standard quality measures. Furthermore, the model's efficacy is validated via a statistical analysis ANOVA test. It is demonstrated that the developed model can be used effectively for early prediction of the substrate structure. The pipeline of this study is flexible and can be extended for in vitro assessment efficacy of anticancer drug response, identification of BCRP functions in transport experiments, and prediction of prostate or lung cancer cell lines.

show abstract

“…Masud et al [ 18 ] leveraged eight different fine-tuned, pre-trained models to classify breast cancers on BUS images and employed a shallow custom convolutional neural network (CNN) for classification. Podda et al [ 19 ] combined several CNNs through specialized ensembles and presented a cyclic mutual optimization step to exploit the intermediate results of the classification in an iterative manner. Jabeen et al [ 20 ] employed deep learning and the fusion of the best selected features for BUS classification, which included data augmentation, pre-trained DarkNet-53 model refining, transfer learning and features extraction, feature selection using two improved optimization algorithms known as reformed differential evaluation (RDE) and reformed gray wolf (RGW), and feature fusion using a probability-based serial approach and classification using machine learning algorithms.…”

Section: Introductionmentioning

confidence: 99%

A Novel Fuzzy Relative-Position-Coding Transformer for Breast Cancer Diagnosis Using Ultrasonography

Guo,

Jiang,

et al. 2023

Healthcare

View full text Add to dashboard Cite

Breast cancer is a leading cause of death in women worldwide, and early detection is crucial for successful treatment. Computer-aided diagnosis (CAD) systems have been developed to assist doctors in identifying breast cancer on ultrasound images. In this paper, we propose a novel fuzzy relative-position-coding (FRPC) Transformer to classify breast ultrasound (BUS) images for breast cancer diagnosis. The proposed FRPC Transformer utilizes the self-attention mechanism of Transformer networks combined with fuzzy relative-position-coding to capture global and local features of the BUS images. The performance of the proposed method is evaluated on one benchmark dataset and compared with those obtained by existing Transformer approaches using various metrics. The experimental outcomes distinctly establish the superiority of the proposed method in achieving elevated levels of accuracy, sensitivity, specificity, and F1 score (all at 90.52%), as well as a heightened area under the receiver operating characteristic (ROC) curve (0.91), surpassing those attained by the original Transformer model (at 89.54%, 89.54%, 89.54%, and 0.89, respectively). Overall, the proposed FRPC Transformer is a promising approach for breast cancer diagnosis. It has potential applications in clinical practice and can contribute to the early detection of breast cancer.

show abstract

Fully-automated deep learning pipeline for segmentation and classification of breast ultrasound images

Cited by 27 publications

References 40 publications

CAM-QUS guided self-tuning modular CNNs with multi-loss functions for fully automated breast lesion classification in ultrasound images

CAM-QUS guided self-tuning modular CNNs with multi-loss functions for fully automated breast lesion classification in ultrasound images

Developing a Breast Cancer Resistance Protein Substrate Prediction System Using Deep Features and LDA

A Novel Fuzzy Relative-Position-Coding Transformer for Breast Cancer Diagnosis Using Ultrasonography

Contact Info

Product

Resources

About