Classification of Histopathological Images of Breast Cancer Tissues into Benign Subcategories and Malignant Subcategories

Abstract

Breast cancer stands as a leading and dangerous illness which affects people throughout the world. Correct diagnosis of breast cancer through histopathological image classification serves as a fundamental step for developing successful treatment strategies. The research introduces a reliable deep-learning system which detects breast cancer tissues between benign and malignant categories for enhanced medical diagnostic precision. The designed model implements state-of-the-art image analytics methods with deep neural networks to obtain valuable information from histopathological pictures. Our techniques (ResNet 50, Inception V3)display better classification accuracy and computing speed in contrast to traditional methods according to comparative assessments. The research findings advance automated histopathological analysis by enabling pathologists to make better clinical decisions through their assistance.

Introduction

1. Introduction

Breast cancer is a leading cause of cancer-related deaths among women globally. Early and accurate diagnosis is critical for successful treatment. Traditional pathology, involving manual examination of tissue samples, is time-consuming and subjective. Recent advances in deep learning, especially using Convolutional Neural Networks (CNNs), offer more precise, faster, and automated image-based diagnosis tools.

2. Related Work

• Traditional ML Techniques like SVM, k-NN, and Random Forest rely on manual feature extraction but struggle with image quality and class imbalance.

• CNNs & Deep Learning have revolutionized histopathological image analysis by automating feature extraction and achieving higher accuracy. Architectures like ResNet, Inception, VGG, and DenseNet are widely used.

• Challenges include class imbalance, interpretability of CNNs (black-box issue), and computational demands.

• Advancements include:

  • Attention mechanisms for focused learning

  • Hybrid CNN-transformer models

  • GANs for data augmentation

• Fine-grained classification (e.g., distinguishing benign subtypes like fibroadenoma vs. malignant types like ductal carcinoma) remains complex due to visual similarity and limited data.

3. Contribution of This Study

• Developed a deep learning pipeline using CNNs (ResNet-50 and InceptionV3) for benign vs. malignant classification.

• Applied advanced preprocessing, data augmentation, and transfer learning to improve model generalization.

• Evaluated performance using standard metrics: accuracy, precision, recall, F1-score, and AUC.

4. Proposed Methodology

• Datasets Used: BreakHis and Kaggle Breast Histopathology dataset.

• Preprocessing:

  • Image resizing, augmentation (rotation, contrast, noise), normalization

  • Data split: 70% training, 15% validation, 15% testing

• Model Architecture:

  • CNN layers (Conv, ReLU, MaxPooling, BatchNorm)

  • Fully connected layers with Softmax for classification

• Transfer Learning:

  • Pre-trained ResNet-50 and InceptionV3

  • Final layers fine-tuned for classification; earlier layers frozen

• Optimization:

  • Adam optimizer, dynamic learning rate scheduling

• Evaluation Metrics:

  • Accuracy, Precision, Recall, F1-score, Error rate

  • ROC-AUC, confusion matrices, and training curves

5. Experimental Setup and Results

• Hardware: RTX 2050 GPU, 16 GB RAM

• Dataset: 211 histopathology images (Benign: 104, Malignant: 107), resized to 224×224 pixels

• Results:

• Key Observations:

  • ResNet-50 outperformed InceptionV3 in all metrics.

  • ROC-AUC: ResNet-50 (0.981), InceptionV3 (0.957)

  • ResNet-50 had better generalization and fewer misclassifications.

  • GLCM analysis showed malignant tissues had higher contrast and dissimilarity, indicating more disordered patterns.

6. Deployment & Tools

• Frameworks: TensorFlow, Keras, OpenCV, scikit-learn

• Tools: TensorBoard, Grad-CAM, Matplotlib for visualization

• Considerations for real-world use:

  • Model compression (quantization/pruning)

  • Web/mobile deployment for pathologist support

7. Conclusion & Future Work

• The study confirms that ResNet-50 provides excellent classification accuracy and robustness for histopathological breast cancer images.

• Future directions include:

  • Larger datasets and diverse subtypes

  • Explainable AI methods to improve interpretability

  • Integration into clinical workflows for real-time diagnostics

  • Exploring nano-feature extraction and hybrid AI models for enhanced precision

Conclusion

This study presents a deep learning-based framework for the classification of histopathological breast cancer images into malignant and benign subtypes. The system incorporates image preprocessing, texture feature extraction using GLCM, and classification using fine-tuned pretrained models—ResNet-50 and InceptionV3. Both models demonstrated strong generalization, achieving a classification accuracy of 95% on the test set. The integration of AnoGAN further enhanced anomaly detection, while GLCM features contributed to improved interpretability and diagnostic robustness. Overall, the proposed system offers a reliable, automated diagnostic aid with significant potential for clinical deployment. Future enhancements may focus on improving accuracy across various magnification levels, automating key diagnostic steps such as ROI detection and report generation, and deploying lightweight models for real-time, edge-based inference. The framework can be extended to incorporate transformer-based deep learning architectures—such as Vision Transformers (ViTs) or Swin Transformers—to capture global contextual relationships within complex tissue structures. Additional directions include integrating 3D histopathological imaging for volumetric analysis and expanding the classification pipeline to other cancer types like lung, prostate, or skin cancer. These advancements will further broaden the applicability and clinical impact of the proposed system in computational pathology.

References

[1] Majid Nawaz, Adel A. Sewissy and Taysir Hassan A. Soliman, “Multi-Class Breast Cancer Classification using Deep Learning Convolutional Neural Network” International Journal of Advanced Computer Science andApplications(ijacsa),9(6),2018. http://dx.doi.org/10.14569/IJACSA.2018.090645 [2] V, L. P. C., G, B. V., R., V. B., & Ramachandran, S. (2024). Deep learning approaches for breast cancer detection in histopathology images: A review. In Cancer Biomarkers (Vols. 40, pp. 1-25). Cancer Biomarkers. https://doi.org/10.3233/CBM-230251 [3] Kanavos, A., Kolovos, E., Papadimitriou, O., & Maragoudakis, M. (2022). Breast Cancer Classification of Histopathological Images using Deep Convolutional Neural Networks. In 2022 7th South-East Europe Design Automation, Computer Engineering, Computer Networks and Social Media Conference (SEEDA-CECNSM) (pp. 1-6). 2022 7th South-East Europe Design Automation, Computer Engineering, Computer Networks and Social Media Conference (SEEDA-CECNSM).https://doi.org/10.1109/SEEDACECNSM57760.2022.9932898 [4] Eltoukhy, M., Hosny, K. M., & Kassem, M. A. (2022). Classification of Multiclass Histopathological Breast Images Using Residual Deep Learning. In Computational Intelligence and Neuroscience (Vols. 2022). Computational Intelligence and Neuroscience. https://doi.org/10.1155/2022/9086060 [5] A, A., Ferlin, D., Shahila, D., Angel, M., Suji, M., Rubia, J., Prathaban, B. P., Vinod, N., & Kumar, . (2024). Deep Learning ResNet-50 Framework for Automatic Early Breast Cancer Diagnosis. In 2024 International Conference on Communication, Computing and Internet of Things (IC3IoT) (pp. 1-5). 2024 International Conference on Communication, Computing and Internet of Things (IC3IoT). https://doi.org/10.1109/IC3IoT60841.2024.10550264 [6] Singh, S., & Kumar, R. (2021). Breast cancer detection from histopathology images with deep inception and residual blocks. In Multimedia Tools and Applications (Vols. 81, pp. 5849-5865). Multimedia Tools and Applications. https://doi.org/10.1007/s11042-021-11775-2 [7] Prasad, C., Arun, B., Amulya, S., Abboju, P., Kollem, S., & Yalabaka, S. (2023). Breast Cancer Classification using CNN with Transfer Learning Models. In 2023 International Conference for Advancement in Technology (ICONAT) (pp. 1-5). 2023 International Conference for Advancement in Technology (ICONAT). https://doi.org/10.1109/ICONAT57137.2023.10080148 [8] Saini, M., & Susan, S. (2019). Data Augmentation of Minority Class with Transfer Learning for Classification of Imbalanced Breast Cancer Dataset Using Inception-V3 (pp. 409-420). https://doi.org/10.1007/978-3-030-31332-6_36 [9] Eskandari, S., Eslamian, A., & Cheng, Q. (2024). Comparative Analysis of Transfer Learning Models for Breast Cancer Classification. In ArXiv: Vol. abs/2408.16859. ArXiv. https://doi.org/10.48550/arXiv.2408.16859 [10] Humayun, M., Khalil, M. I., Almuayqil, S., & Jhanjhi, N. Z. (2023). Framework for Detecting Breast Cancer Risk Presence Using Deep Learning. In Electronics. Electronics. https://doi.org/10.3390/electronics12020403. [11] Cai, X., Li, X., Razmjooy, N., & Ghadimi, N. (2021). Breast cancer diagnosis by convolutional neural network and advanced thermal exchange optimization algorithm. Computational and Mathematical Methods in Medicine, 2021(1), 5595180. [12] Khan, S., Islam, N., Jan, Z., Din, I. U., & Rodrigues, J. J. C. (2019). A novel deep learning based framework for the detection and classification of breast cancer using transfer learning. Pattern Recognition Letters, 125, 1-6. [13] Yusoff, M., Haryanto, T., Suhartanto, H., Mustafa, W. A., Zain, J. M., & Kusmardi, K. (2023). Accuracy Analysis of Deep Learning Methods in Breast Cancer Classification: A Structured Review. In Diagnostics (Vols. 13). Diagnostics. https://doi.org/10.3390/diagnostics13040683 [14] Ahmad, J., Akram, S., Jaffar, A., Ali, Z., Bhatti, S., Ahmad, A., & Rehman, S. ur. (2024). Deep learning empowered breast cancer diagnosis: Advancements in detection and classification. In PLOS ONE (Vols. 19). PLOS ONE. https://doi.org/10.1371/journal.pone.0304757 [15] Aldakhil, L. A., Alhasson, H. F., & Alharbi, S. S. (2024). Attention-Based Deep Learning Approach for Breast Cancer Histopathological Image Multi-Classification. In Diagnostics (Vols. 14). Diagnostics. https://doi.org/10.3390/diagnostics14131402 [16] Nawaz, M., Sewissy, A. A., & Soliman, T. H. A. (2018). Multi-class breast cancer classification using deep learning convolutional neural network. Int. J. Adv. Comput. Sci. Appl, 9(6), 316-332. [17] Abhisheka, B., Biswas, S., & Purkayastha, B. (2023). A Comprehensive Review on Breast Cancer Detection, Classification and Segmentation Using Deep Learning. In Archives of Computational Methods in Engineering (Vols. 30, pp. 5023-5052). Archives of Computational Methods in Engineering. https://doi.org/10.1007/s11831-023-09968-z [18] M. G. Mohiuddin, D. L. S. Reddy and A. Ahmed, \"Detection of Cancer Cells Using Convolutional Neural Networks (CNNs),\" 2024 2nd International Conference on Computing and Data Analytics (ICCDA), Shinas, Oman, 2024, pp. 1-6, doi: 10.1109/ICCDA64887.2024.10867378 https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10867378&isnumber=10867267 [19] Varavadekar, Varad & Sajwan, Vineet & Mazo, Claudia. (2024). Transfer Learning Approach for Detecting Breast Cancer Tumour Subtypes in Whole-SlideImages.10.1109/ISBI56570.2024.10635534.https://www.researchgate.net/publication/382870622_Transfer_Learning_Approach_for_Detecting_Breast_Cancer_Tumour_Subtypes_in_Whole-Slide_Images [20] Dubey, A., Yadav, P., & Bhargava, C. P. (2024). Classification of Breast Cancer Histopathological Images Using Residual Learning-based CNN. https://koreascience.kr/article/JAKO202407664805721.pdf [21] Xiao, Ting, Liu, Lei, Li, Kai, Qin, Wenjian, Yu, Shaode, Li, Zhicheng, Comparison of Transferred Deep Neural Networks in Ultrasonic Breast Masses Discrimination, BioMed Research International, 2018, 4605191, 9 pages, 2018. https://doi.org/10.1155/2018/4605191

Copyright

Copyright © 2025 Asst. Prof. D.V.N. Bharathi, Bodapati Subrahmanyam, Chollangi Shanmukha Venkata Sai, Addala Sai Praneeth, Durgasi Praveen. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.