Asisten Kanker Payudara

Portal Penelitian CNN Breast Cancer

No Type BC Dataset Model Overview of CNN Blocks Result Author Keterangan
1 Mitotic cell detection ICPR 2012 Mitosis Dataset and ICPR 2014 Mitosis Dataset AlexNet untuk klasifikasi berbasis patch dengan preprocessing, dan U-Net untuk segmentasi semantik end-to-end Not reported Acc=0.9419, F1-score=0.9435, Sens=0.9419, Spec=0.9577, Dice=0.9747 Deep Learning for Semantic Segmentation vs. Classification in Computational Pathology Application to Mitosis Analysis in Breast Cancer Grading
2 Breast Pathology Wisconsin breast cancer (WBC) Support Vector Machine (SVM) Not reported The study confirms that SVM with image preprocessing and feature extraction enables accurate breast cancer classification, offering potential for early diagnosis. Detection of Breast Cancer Using Machine Learning Support Vector Machine Algorithm
3 Mass Lesions BC The dataset comprised 400 mammogram cases from women in northeast China. CNN extracts features, while ELM performs clustering and classification on fused deep features. Not reported Multi-Feature Model (MF) dengan CNN deep feature + Morphological feature + Texture feature + Density feature (CNN-GTD): Acc=0.8650, Sens=0.8510, Spec=0.8820, AUC=0.9230 Breast Cancer Detection Using Extreme Learning Machine Based on Feature Fusion With CNN Deep Features
4 Malign Breast Cancer The combined clinical cases from multiple datasets, encompassing both screening and clinical data, amounted to a total of 2,652 cases collected from Sweden, the UK, the Netherlands, Italy, and Spain. The study employed the Transpara v1.4.0 AI system, a deep learning?based CNN model developed by ScreenPoint Medical BV. Not reported AUC=0.840 (95% CI: 0.820 sampai 0.860) Stand-Alone Artificial Intelligence for Breast Cancer Detection in Mammography Comparison With 101 Radiologists
5 Mitotic cell detection ICPR 2014 MITOS-ATYPIA-14 Dataset, ICPR 2012 MITOSIS Dataset, and TUPAC16 (Tumor Proliferation Assessment Challenge 2016) dataset. TL-HCNN-Mit-Det (Transfer Learning and Hybrid CNN based Mitosis Detection) Reported Prec=0.772, Rec=0.663, F-measure=0.713 Transfer learning based deep CNN for segmentation and detection of mitoses in breast cancer histopathological images
6 Breast Pathology Mammographic Image Analysis Society (MIAS) K-means clustering ollowed by classification employing a Multiclass Support Vector Machine (MSVM) Reported The proposed DL method achieved mean Acc=0.95 (normal), 0.94 (benign), and 0.98 (malignant), with an overall 2% improvement over prior state-of-the-art methods. MSVM: Acc=0969, outperforming KNN=0.938, LDA=0.897, and DT=0.887, while SVM yielded an AUC=0.99; results were validated via 10-fold cross-validation with a 70/30 train-test split. Intellectual detection and validation of automated mammogram breast cancer images by multi-class SVM using deep learning classification
7 Malign Breast Cancer The dataset, comprising 3,002 mammogram images, was collected from the Department of Breast and Endocrine Surgery at Hallym University Sacred Heart Hospital between February 2007 and May 2015. Convolutional Neural Network (CNN) Reported DenseNet-169 and EfficientNet-B5 both performed excellently in breast cancer detection, with DenseNet-169 achieving AUC 0.952 ? 0.005 (sensitivity 87.0%, specificity 88.4%) and EfficientNet-B5 slightly higher AUC 0.954 ? 0.020, though DenseNet-169 showed greater stability. Automated Breast Cancer Detection in Digital Mammograms of Various Densities via Deep Learning
8 Mitotic cell detection ICPR 2014 MITOS-ATYPIA-14 Dataset, ICPR 2012 MITOSIS Dataset, and TUPAC16 (Tumor Proliferation Assessment Challenge 2016) dataset. Faster R-CNN + Post Processing + Score Level Fusion Reported On the ICPR 2012 dataset, the model achieved prec=0.876, rec=0.841, F1-score=0.858, whereas on the ICPR 2014 dataset, it achieved prec=0.848, recall=0.583, and F1-score=0.691. Artificial Intelligence-Based Mitosis Detection in Breast Cancer Histopathology Images Using Faster R-CNN and Deep CNNs
9 Mass Lesions BC The dataset used in this study comprises 55 anonymized DCE?MRI scans of breast cancer patients, including 11 cases of DCIS + LCIS and 44 cases of IDC. The proposed model combines a pretrained CNN (GoogleNet) for feature extraction, followed by an ANN for false-positive reduction in detection, and a Radiomics-based classifier for tumor classification. Reported F1-score=0.9000, Sens=0.7500, AUC=0.7000 Breast Cancer Mass Detection in DCE-MRI Using Deep-Learning Features Followed by Discrimination of Infiltrative vs. In Situ Carcinoma through a Machine-Learning Approach
10 Breast Pathology Wisconsin breast cancer (WBC), Breast Cancer Wisconsin (Diagnostic) Dataset, Breast Cancer Wisconsin (Prognostic) Dataset, Manuel Gomes from the University Hospital Centre of Coimbra. The final model integrates Principal Component Analysis (PCA) for dimensionality reduction, a Multilayer Perceptron (MLP) for feature extraction, and a Support Vector Machine (SVM) as the classifier through transfer learning. Reported PCA reduced data dimensionality by 99.89%, while MLP effectively extracted features. The MLP2SVM model achieved 86.97% accuracy under 10-fold cross-validation, showing a minimal 0.38% accuracy gap but up to 43 times higher efficiency than MLP2DT, confirming SVM as the most suitable classifier. Breast Cancer-Detection System Using PCA, Multilayer Perceptron, Transfer Learning, and Support Vector Machine

📈 Tren Penelitian Berdasarkan Jenis (Type BC)