RECORD
Application of Transfer Learning and Ensemble Learning in Image-level Classification for Breast Histopathology
Computer Vision and Pattern Recognition · Submitted 18 Apr 2022
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim Our model has outperformed many existing approaches in accuracy, providing a method for the field of auxiliary medical diagnosis.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
Background: Breast cancer has the highest prevalence in women globally. The classification and diagnosis of breast cancer and its histopathological images have always been a hot spot of clinical concern. In Computer-Aided Diagnosis (CAD), traditional classification models mostly use a single network to extract features, which has significant limitations. On the other hand, many networks are trained and optimized on patient-level datasets, ignoring the application of lower-level data labels. Method: This paper proposes a deep ensemble model based on image-level labels for the binary classification of benign and malignant lesions of breast histopathological images. First, the BreaKHis dataset is randomly divided into a training, validation and test set. Then, data augmentation techniques are used to balance the number of benign and malignant samples. Thirdly, considering the performance of transfer learning and the complementarity between each network, VGG16, Xception, ResNet50, DenseNet201 are selected as the base classifiers. Result: In the ensemble network model with accuracy as the weight, the image-level binary classification achieves an accuracy of $98.90\%$. In order to verify the capabilities of our method, the latest Transformer and Multilayer Perception (MLP) models have been experimentally compared on the same dataset. Our model wins with a $5\%-20\%$ advantage, emphasizing the ensemble model's far-reaching significance in classification tasks. Conclusion: This research focuses on improving the model's classification performance with an ensemble algorithm. Transfer learning plays an essential role in small datasets, improving training speed and accuracy. Our model has outperformed many existing approaches in accuracy, providing a method for the field of auxiliary medical diagnosis.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | Background: Breast cancer has the highest prevalence in women globally. | background 1.00 | 0.03 | 0.01 | asserted 1.00 |
| s1 | The classification and diagnosis of breast cancer and its histopathological images have always been a hot spot of clinical concern. | background 1.00 | 0.04 | 0.01 | asserted 1.00 |
| s2 | In Computer-Aided Diagnosis (CAD), traditional classification models mostly use a single network to extract features, which has significant limitations. | background 0.72 | 0.04 | 0.02 | asserted 1.00 |
| s3 | On the other hand, many networks are trained and optimized on patient-level datasets, ignoring the application of lower-level data labels. | gap 0.73 | 0.05 | 0.01 | asserted 1.00 |
| s4 | Method: This paper proposes a deep ensemble model based on image-level labels for the binary classification of benign and malignant lesions of breast histopathological images. | method 0.52 | 0.80 | 0.03 | asserted 0.98 |
| s5 | First, the BreaKHis dataset is randomly divided into a training, validation and test set. | method 1.00 | 0.94 | 0.01 | asserted 0.98 |
| s6 | Then, data augmentation techniques are used to balance the number of benign and malignant samples. | method 1.00 | 0.93 | 0.01 | asserted 0.97 |
| s7 | Thirdly, considering the performance of transfer learning and the complementarity between each network, VGG16, Xception, ResNet50, DenseNet201 are selected as the base classifiers. | method 1.00 | 0.88 | 0.02 | asserted 0.91 |
| s8 | Result: In the ensemble network model with accuracy as the weight, the image-level binary classification achieves an accuracy of $98.90\%$. | result 1.00 | 0.97 | 0.18 | asserted 1.00 |
| s9 | In order to verify the capabilities of our method, the latest Transformer and Multilayer Perception (MLP) models have been experimentally compared on the same dataset. | method 0.81 | 0.91 | 0.02 | asserted 0.94 |
| s10 | Our model wins with a $5\%-20\%$ advantage, emphasizing the ensemble model's far-reaching significance in classification tasks. | interpretation 0.60 | 0.93 | 0.12 | asserted 1.00 |
| s11 | Conclusion: This research focuses on improving the model's classification performance with an ensemble algorithm. | contribution 0.88 | 0.63 | 0.88 | asserted 0.91 |
| s12 | Transfer learning plays an essential role in small datasets, improving training speed and accuracy. | interpretation 0.82 | 0.20 | 0.08 | asserted 1.00 |
| s13 | Our model has outperformed many existing approaches in accuracy, providing a method for the field of auxiliary medical diagnosis. | interpretation 0.52 | 0.85 | 0.28 | asserted 1.00 |
What was examined, and what was not
- arXiv listing abstract v2
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
Background: Breast cancer has the highest prevalence in women globally. The classification and diagnosis of breast cancer and its histopathological images have always been a hot spot of clinical concern. In Computer-Aided Diagnosis (CAD), traditional classification models mostly use a single network to extract features, which has significant limitations. On the other hand, many networks are trained and optimized on patient-level datasets, ignoring the application of lower-level data labels. Method: This paper proposes a deep ensemble model based on image-level labels for the binary classification of benign and malignant lesions of breast histopathological images. First, the BreaKHis dataset is randomly divided into a training, validation and test set. Then, data augmentation techniques are used to balance the number of benign and malignant samples. Thirdly, considering the performance of transfer learning and the complementarity between each network, VGG16, Xception, ResNet50, DenseNet201 are selected as the base classifiers. Result: In the ensemble network model with accuracy as the weight, the image-level binary classification achieves an accuracy of $98.90\%$. In order to verify the capabilities of our method, the latest Transformer and Multilayer Perception (MLP) models have been experimentally compared on the same dataset. Our model wins with a $5\%-20\%$ advantage, emphasizing the ensemble model's far-reaching significance in classification tasks. Conclusion: This research focuses on improving the model's classification performance with an ensemble algorithm. Transfer learning plays an essential role in small datasets, improving training speed and accuracy. Our model has outperformed many existing approaches in accuracy, providing a method for the field of auxiliary medical diagnosis.
Human review
How the literature uses it
What this paper does with the arXiv papers it cites
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.