Breast cancer remains the most prevalent malignancy among women worldwide. The timely detection of this cancer type is critical for improving survival outcomes. Despite advancements, mammogram classification using deep learning strategies still faces challenges. These include inter-view feature inconsistency, loss of diagnostic details, and limited interpretability. In order to address these issues, MammoFusion-Net, a dual-branch deep learning framework, is proposed for mammogram-based breast cancer classification. Using residual convolutional streams, the framework processes craniocaudal (CC) and mediolateral oblique (MLO) views independently. This supports preservation of view-specific anatomical information. In the proposed framework, a Gates Cross-View Fusion mechanism adaptively integrates features across views. As a result of experimental analysis, the proposed framework achieved 92.116 % (VinDr-Mammo dataset) and 95.556 % (INBreast dataset) of improved classification performance.•Employs a dual-branch architecture to independently process CC and MLO views using residual convolutional streams.•Integrates Gated Cross-View Fusion and attention mechanisms adaptively and refines multi-view features for stronger discrimination.•Demonstrates the explainability of the model through Grad-CAM visualizations that highlight lesion-relevant regions.

MLSE-FML: Mammogram classification with multi-level semantic extraction and focal margin loss

This study presents a YOLO (You Only Look Once)-based approach for the automated detection of breast lesions in mammography images; the corresponding implementation is provided in the Appendix. The proposed method adopts a continuous training scheme across three public mammography datasets, organized according to increasing annotation quality and domain complexity. Initially, the model is trained on the CBIS-DDSM dataset, which consists of digitized film mammograms with coarse annotations, achieving an mAP@0.5 of 45.6% and a precision of 50.7%. Subsequently, the resulting model is fine-tuned on the INBreast dataset, which comprises high-quality digital mammograms with expert-level annotations, leading to substantial performance improvements, with an mAP@0.5 of 76.73% and a precision of 83.32%. Finally, the refined model is further fine-tuned on the VinDr-Mammo dataset, yielding additional gains and reaching an mAP@0.5 of 80.55% and a precision of 80.07%. Overall, the results demonstrate consistent improvements in accuracy and robustness across heterogeneous mammographic datasets.

This paper presents a novel deep learning framework for accurate and interpretable breast cancer detection using multi-view mammogram images. The architecture integrates a cross-attention transformer for global relational learning between craniocaudal (CC) and mediolateral oblique (MLO) views, and a graph neural network (GNN) to model intra-view region-level structural topology dependencies. A prototype reasoning module enables explainable classification by comparing extracted features with learned benign and malignant exemplars. To enhance model robustness and hyperparameter tuning, a Hybrid Harris Hawks Optimization–Slime Mould Algorithm (HHO–SMA) is employed to jointly optimize attention weights, prototype thresholds, and fusion parameters. The proposed method is evaluated on the INbreast dataset and demonstrates superior accuracy, sensitivity, and explainability compared to state-of-the-art convolutional neural network and transformer baselines. The incorporation of HHO–SMA significantly improves optimization convergence, ensuring high diagnostic reliability in clinical settings.

Early detection of breast cancer by mammography scans is crucial for improving treatment outcomes. However, low image resolutions, size, and location of lesions in dense breast tissue prove to be challenges in mammography, underscoring the importance of accurate and efficient computer-aided diagnostic systems. This article introduces a novel classification framework that utilizes histogram of oriented gradients (HOG) as a feature extractor and principal component analysis (PCA) for dimensionality reduction. Classification is implemented using the persistent homology classification algorithm (PHCA), which leverages persistent homology (PH) to capture topological properties of mammography images. The framework was evaluated on 7,632 images from the INbreast dataset with an extensive use of grid-search cross-validation to optimize preprocessing parameters. Two optimal combinations of HOG parameters and scaler were identified, with the best configuration (16 × 16 pixels per cell, 3 × 3 cells per block, and Minmax scaler) achieving strong performance. Validating on the test set, PHCA achieved an overall accuracy, precision, recall, F1-score, and specificity of 97.31%, 96.86%, 97.09%, 96.97%, and 96.86%, respectively. Clinically, the high precision (98.23%) and high recall (97.75%) for malignant cases highlight PHCA’s sensitivity in identifying malignancies, ensuring that very few malignant cases go undetected with highly trustworthy predictions. These results are shown to be competitive with existing state-of-the-art models, even exceeding in some cases, while requiring lower computational cost than deep learning-based approaches. Although the proposed method trails advanced deep models by 3–4% in some metrics, it offers a computationally efficient alternative and a potential for deployment in large-scale screening systems, demonstrating the promise of topological data analysis for breast cancer classification.

Purpose: Manually segmenting mammograms is time-consuming and subjective. Therefore, automatic segmentation of breast masses is necessary but poses significant challenges due to factors such as low signal-to-noise ratio, diverse mass shapes and sizes, varying contrast levels, and high false positive rates. To address these challenges, we have developed an automatic image segmentation method based on a comprehensive pre-processing pipeline. Materials and Methods: Our proposed method consists of two phases: 1) the pre-processing phase, which includes denoising, contrast enhancement, image cropping, resizing, and augmentation of mammograms, and 2) the model design phase, where UNet++ is employed as an encoder-decoder-based network for segmenting breast masses. The encoder captures relevant information from various regions in the input image, while the decoder reconstructs the spatial location of the target region. We conducted extensive experiments on publicly available CBIS-DDSM and INbreast datasets to evaluate the performance of our proposed method. For a comprehensive assessment, we utilized evaluation metrics including Precision, True Positive Rate, Dice Score Coefficient, and Jaccard Index. Additionally, a confusion matrix was employed to evaluate segmentation accuracy, while violin plots depicted the distribution of results across different BI-RADS and ACR categories. Results: Based on our findings, our proposed method demonstrates promising results with a precision rate of 92.33%, a True Positive Rate of 93.83%, a Dice Score Coefficient measuring 92.92%, and a Jaccard Index of 87.05% in the CBIS-DDSM dataset. Furthermore, to assess the generalizability of our proposed method, the INbreast dataset was used as an unseen test set. The results demonstrate a precision rate of 91.15%, a True positive rate of 91.15%, a Dice Score coefficient of 92.53%, and a Jaccard Index of 87.25%, indicating robust performance on data outside the training distribution. Conclusion: The integration of UNet++ with a pre-processing pipeline in digital mammography has shown promising results in accurately segmenting breast masses. This method has the potential to significantly improve early breast cancer detection and reduce diagnostic errors in clinical practice while employing a relatively lightweight model.

Medical imaging techniques like mammography enable early breast cancer detection and are part of regular screening programs. Typically, a mammogram exam involves two views of each breast, providing complementary information, but physicians rate the breast as a whole. Computer-Aided Diagnostic tools focus on detecting lesions in a single view, which is challenging due to high image resolution and varying scales of abnormalities. The projective nature of the two views and different acquisition protocols add complexity to dual-view analysis. To address these challenges, we propose a Graph Neural Network approach that models image information at multiple scales and the complementarity of the two views. To this end, we rely on a superpixel decomposition, assigning hierarchical features to superpixels, designing a dual-view graph to share information, and introducing a modified Sparse Graph Attention Layer to keep relevant dual-view relations. This improves interpretability of decisions and avoids the need to register pairs of views under strong deformations. Our model is trained with a fully supervised approach and evaluated on public and private datasets. Experiments demonstrate state-of-the-art classification and detection performance on Full Field Digital Mammographies, achieving a breast-wise AUC of 0.96 for the INbreast dataset, a sensitivity of 0.97 with few false positives per image (0.33), and a case-wise AUC of 0.92 for the VinDr dataset. This study presents a Sparse Graph Attention method for dual-view mammography analysis, generating meaningful explanations that radiologists can interpret. Extensive evaluation shows the relevance of our approach in breast cancer detection and classification.

Abstract Breast cancer is one of the leading causes of cancer-related morbidity worldwide, underscoring the need for advanced diagnostic tools to improve early detection and treatment outcomes. This study introduces MammoSegNet, a novel convolutional neural network architecture optimized for precisely segmenting mammographic images. The proposed MammoSegNet incorporates Inception-ResNet blocks, Squeeze-and-Excitation (SE) modules, and dilated convolutions to enable multi-scale feature extraction and efficient attention refinement while maintaining low computational complexity. MammoSegNet performance was rigorously evaluated on BCDR-D01 and INbreast datasets to examine its robustness and generalization. Using stratified fivefold cross-validation, the model was trained on BCDR-D01 and tested on the unseen INbreast dataset through Monte Carlo cross-validation. Preprocessing techniques, including Region of Interest (ROI) Isolation to concentrate on relevant areas, Normalization to standardized pixel intensities, and Data Augmentation to expand the dataset and enhance the model’s robustness, were employed. Additionally, a specialized image enhancement method called peak feature intensity transformation (PFIT) was designed to amplify diagnostic features while preserving structural integrity. Comparative evaluations confirmed MammoSegNet’s superior performance across metrics, achieving 97% accuracy on BCDR-D01 and 95% on INbreast. Statistical t-tests validated these improvements, and visual heatmaps demonstrated the model’s effectiveness in isolating tumor regions. These findings establish MammoSegNet as a promising tool for enhancing breast cancer diagnostic accuracy and reliability in medical applications.

ABSTRACT This study introduces a novel hybrid deep learning model that combines residual convolutional networks and a multilayer perceptron (MLP)‐based transformer for precise breast lesion segmentation and classification using mammogram images. Initially, mammograms undergo preprocessing involving thresholding and Gabor‐based pixel segmentation to extract informative patches. The proposed model leverages deep features extracted via convolutional neural networks, which are subsequently processed through self‐attention and cross‐attention mechanisms in a modified transformer architecture to capture both local and global dependencies for classification. The approach is rigorously evaluated on the publicly available INbreast dataset, achieving classification accuracies of 98.17% for a three‐class (normal, benign, malignant) scenario and 96.74% for a more detailed five‐class classification. The model demonstrates strong capabilities in differentiating subtle variations between malignant and benign tissues. These promising results suggest significant potential for practical clinical implementation, assisting radiologists by providing highly accurate diagnostic insights. Notably, this approach contributes substantially to automated breast cancer diagnostics, highlighting the efficacy of integrating convolutional neural network features with transformer architectures for improved segmentation and classification outcomes.

Breast cancer is a leading cause of mortality among women worldwide, and early detection is critical for improving survival rates. While mammography is a key screening tool, its accuracy can be impacted by human interpretation. Convolutional Neural Networks (CNNs) offer advanced image analysis capabilities to enhance early detection and support healthcare professionals with higher accuracy and reliability. This study presents a novel CNN architecture, developed from scratch, to automate breast cancer detection and improve diagnostic accuracy. Using the MIAS and INBREAST datasets with advanced data augmentation techniques, the model demonstrates outstanding performance. On the MIAS dataset, it achieves an accuracy of 0.9912, recall of 0.9912, precision of 0.9914, AUC of 0.9996, and F1-score of 0.9912. Similarly, on the INBREAST dataset, the model achieves an accuracy of 0.9494, recall of 0.9494, precision of 0.9529, AUC of 0.9937, and F1-score of 0.9493, highlighting the accuracy and reliability across different datasets. The findings illustrate the potential of deep learning-based computer-aided diagnostic (CAD) systems in improving early breast cancer detection, reducing errors, and enhancing the cost-efficiency of providing healthcare.

Enhancing radiomics features via a large language model for classifying benign and malignant breast tumors in mammography.

.PurposeThe scarcity of high-quality curated labeled medical training data remains one of the major limitations in applying artificial intelligence systems to breast cancer diagnosis. Deep models for mammogram analysis and mass (or micro-calcification) detection require training with a large volume of labeled images, which are often expensive and time-consuming to collect. To reduce this challenge, we proposed a method that leverages self-supervised learning (SSL) and a deep hybrid model, named HybMNet, which combines local self-attention and fine-grained feature extraction to enhance breast cancer detection on screening mammograms.ApproachOur method employs a two-stage learning process: (1) SSL pretraining: We utilize Efficient Self-Supervised Vision Transformers, an SSL technique, to pretrain a Swin Transformer (Swin-T) using a limited set of mammograms. The pretrained Swin-T then serves as the backbone for the downstream task. (2) Downstream training: The proposed HybMNet combines the Swin-T backbone with a convolutional neural network (CNN)-based network and a fusion strategy. The Swin-T employs local self-attention to identify informative patch regions from the high-resolution mammogram, whereas the CNN-based network extracts fine-grained local features from the selected patches. A fusion module then integrates global and local information from both networks to generate robust predictions. The HybMNet is trained end-to-end, with the loss function combining the outputs of the Swin-T and CNN modules to optimize feature extraction and classification performance.ResultsThe proposed method was evaluated for its ability to detect breast cancer by distinguishing between benign (normal) and malignant mammograms. Leveraging SSL pretraining and the HybMNet model, it achieved an area under the ROC curve of 0.864 (95% CI: 0.852, 0.875) on the Chinese Mammogram Database (CMMD) dataset and 0.889 (95% CI: 0.875, 0.903) on the INbreast dataset, highlighting its effectiveness.ConclusionsThe quantitative results highlight the effectiveness of our proposed HybMNet and the SSL pretraining approach. In addition, visualizations of the selected region of interest patches show the model’s potential for weakly supervised detection of microcalcifications, despite being trained using only image-level labels.

This study introduces a novel computer aided diagnosis system to diagnose breast cancer using two mammography views as input i.e. MLO and CC. The pipeline consists of a convolutional autoencoder that is trained to extract features from different mammograms’ views, and one-dimensional convolutional neural nework to classify the input embeddings into two classes i.e. benign or malignant. We compare the one-dimensional convolutional neural network classification results with a support vector machine trained on the same latent embeddings. We conclude that the combination of autoencoders and one-dimensional convolutional neural networks yield the best classification accuracy on the test set of the INbreast dataset.

ABSTRACT Addressing the complexities of classifying distinct object classes in computer vision presents several challenges, including effectively capturing features such as color, form, and tissue size for each class, correlating class vulnerabilities, singly capturing features, and predicting class labels accurately. To tackle these issues, we introduce a novel hybrid deep dense learning technique that combines deep transfer learning with a transformer architecture. Our approach utilizes ResNet50, EfficientNetB1, and our proposed ProDense block as the backbone models. By integrating the Vit‐L16 transformer, we can focus on relevant features in mammography and extract high‐value pair features, offering two alternative methods for feature extraction. This allows our model to adaptively shift the region of interest towards the class type in slides. The transformer architecture, particularly Vit‐L16, enhances feature extraction by efficiently capturing long‐range dependencies in the data, enabling the model to better understand the context and relationships between features. This aids in more accurate classification, especially when fine‐tuning pretrained models, as it helps the model adapt to specific characteristics of the target dataset while retaining valuable information learned from the pretraining phase. Furthermore, we employ a stack ensemble technique to leverage both the deep transfer learning model and the ProDense block extension for training extensive features for breast cancer classification. The fine‐tuning process employed by our hybrid model helps refine the dense layers, enhancing classification accuracy. Evaluating our method on the INbreast dataset, we observe a significant improvement in predicting the binary cancer category, outperforming the current state‐of‐the‐art classifier by 98.08% in terms of accuracy.

The early and accurate detection of breast lesions through mammography is crucial for improving survival rates. However, the existing deep learning-based methods often rely on costly pixel-level annotations, limiting their scalability in real-world applications. To address this issue, a novel local extremum mapping (LEM) mechanism is proposed for mammogram classification and weakly supervised lesion localization. The proposed method first divides the input mammogram into multiple regions and generates score maps through convolutional neural networks. Then, it identifies the most informative regions by filtering local extrema in the score maps and aggregating their scores for final classification. This strategy enables lesion localization with only image-level labels, significantly reducing annotation costs. Experiments on two public mammography datasets, CBIS-DDSM and INbreast, demonstrate that the proposed method achieves competitive performance. On the INbreast dataset, LEM improves classification accuracy to 96.3% with an AUC of 0.976. Furthermore, the proposed method effectively localizes lesions with a dice similarity coefficient of 0.37, outperforming Grad-CAM and other baseline approaches. These results highlight the practical significance and potential clinical applications of our approach, making automated mammogram analysis more accessible and efficient.

Breast cancer has a high incidence and mortality rate in the female population. Mammography is the most reliable method for early and accurate diagnosis of breast cancer. Automated detection and classification of breast masses on mammograms is a challenging task and is essential to assist radiologists in accurately diagnosing breast masses. The aim of this study is to develop a Computer-Aided Diagnosis (CAD) system based on You Look Only Once (YOLO) for identifying breast masses and classifying them as benign or malignant. We propose a YOLOv5-CAD framework that uses a transfer learning approach. Two datasets, CBIS-DDSM and VinDr-Mammo, are utilized for training from scratch. The model weights and parameters are subsequently transferred and fine-tuned onto the smaller INBreast dataset. Furthermore, an analysis is conducted to assess the impact of various data augmentation techniques during the training phase on enhancing model performance. The proposed framework demonstrates encouraging fivefold cross-validation evaluation results. To conclude, transfer learning from CBIS-DDSM achieves 0.843 mAP, precision of 0.855, recall of 0.774, while transfer learning from VinDr- Mammo reaches 0.84 mAP, precision of 0.829, recall of 0.787. Furthermore, the performance of the two fine-tuned models was tested on both the MIAS dataset and the private dataset from Başkent University Ankara Hospital. Such promising performance could be useful for the CAD frameworks being developed to support radiologists as a second opinion reader for the detection and classification of breast masses.

Ensuring strict medical data privacy standards while delivering efficient and accurate breast cancer segmentation is a critical challenge. This paper addresses this challenge by proposing a lightweight solution capable of running directly in the user’s browser, ensuring that medical data never leave the user’s computer. Our proposed solution consists of a two-stage model: the pre-trained nano YoloV5 variation handles the task of mass detection, while a lightweight neural network model of just 20k parameters and an inference time of 21 ms per image addresses the segmentation problem. This highly efficient model in terms of inference speed and memory consumption was created by combining well-known techniques, such as the SegNet architecture and depthwise separable convolutions. The detection model manages an mAP@50 equal to 50.3% on the CBIS-DDSM dataset and 68.2% on the INbreast dataset. Despite its size, our segmentation model produces high-performance levels on the CBIS-DDSM (81.0% IoU, 89.4% Dice) and INbreast (77.3% IoU, 87.0% Dice) dataset.

PurposeBreast cancer is the most common major public health problems of women in the world. Until now, analyzing mammogram images is still the main method used by doctors to diagnose and detect breast cancers. However, this process usually depends on the experience of radiologists and is always very time consuming.Patients and MethodsWe propose to introduce deep learning technology into the process for the facilitation of computer-aided diagnosis (CAD), and address the challenges of class imbalance, enhance the detection of small masses and multiple targets, and reduce false positives and negatives in mammogram analysis. Therefore, we adopted and enhanced RetinaNet to detect masses in mammogram images. Specifically, we introduced a novel modification to the network structure, where the feature map M5 is processed by the ReLU function prior to the original convolution kernel. This strategic adjustment was designed to prevent the loss of resolution for small mass features. Additionally, we introduced transfer learning techniques into training process through leveraging pre-trained weights from other RetinaNet applications, and fine-tuned our improved model using the INbreast dataset.ResultsThe aforementioned innovations facilitate superior performance of the enhanced RetiaNet model on the public dataset INbreast, as evidenced by a mAP (mean average precision) of 1.0000 and TPR (true positive rate) of 1.00 at 0.00 FPPI (false positive per image) on the INbreast dataset.ConclusionThe experimental results demonstrate that our enhanced RetinaNet model defeats the existing models by having more generalization performance than other published studies, and it can also be applied to other types of patients to assist doctors in making a proper diagnosis.

Introduction: Breast Cancer (BC) is a significant cause of high mortality amongst women globally and probably will remain a disease posing challenges about its detectability. Advancements in medical imaging technology have improved the accuracy and efficiency of breast cancer classification. However, tumor features' complexity and imaging data variability still pose challenges. Method: This study proposes the Ensemble Residual-VGG-16 model as a novel combination of the Deep Residual Network (DRN) and VGG-16 architecture. This model is purposely engineered with maximal precision for the task of breast cancer diagnosis based on mammography images. We assessed its performance by accuracy, recall, precision, and the F1-Score. All these metrics indicated the high performance of this Residual-VGG-16 model. The diagnostic residual-VGG16 performed exceptionally well with an accuracy of 99.6%, precision of 99.4%, recall of 99.7%, F1 score of 98.6%, and Mean Intersection over Union (MIoU) of 99.8% with MIAS datasets. Result: Similarly, the INBreast dataset achieved an accuracy of 93.8%, a precision of 94.2%, a recall of 94.5%, and an F1-score of 93.4%. Conclusion: The proposed model is a significant advancement in breast cancer diagnosis, with high accuracy and potential as an automated grading.

One of the most common cancers among women worldwide is breast cancer (BC), and early diagnosis can save lives. Early detection of BC increases the likelihood of a successful outcome by enabling treatment to start sooner. Even in areas without access to a specialist physician, machine learning (ML) aids in early BC detection. The medical imaging community is becoming more interested in using ML, and deep learning (DL) to increase the accuracy of cancer screening. Many disease-related data are sparse. However, for DL models to perform well, a large amount of data is required. Because of this, the DL models that are currently in use on medical images are not as effective as they could be. Convolutional neural network (CNN) models have recently gained popularity in the medical industry, and they perform admirably in terms of high performance and robustness at image classification. The proposed method classifies data using ensemble pre-trained models such as the dense convolutional network (DenseNet)-121 and EfficientNet-B5 feature extractor networks, as well as the support vector machine for classification. Using a modified meta-heuristic optimizer, the selected pre-trained CNN hyperparameters were optimized to improve the performance. The experimental results for the presented model on the INbreast dataset show that the EfficientNet-B5 model is effective for BC classification, with overall accuracy, sensitivity, specificity, precision, and area under the ROC curve (AUC) values of 99.9%, 99.9%, 99.8%, 99.1%, 1.0, respectively.