The article proposes an approach for detecting breast cancer with the aid of machine learning. The main motive of this article is to detect cancer at earlier stage using deep learning (DL) where the ultrasound images of breast are classified as cancerous or non-cancerous. Using CNN model the system is trained to classify based on tumour size. These ultrasound images are pre-processed using normalization and are resized beforehand so that the learned model can make predictions with ease. The system even detects the stage of cancer if the tumour is predicted to be cancerous using prediction probability. Overall this helps the medical professionals to detect breast cancer easily and provide proper medication to the affected ones and prevent human error.
Introduction
Breast cancer is a leading cause of death worldwide, accounting for nearly 25% of cancer cases, with incidence rising annually. Early detection is critical for effective treatment and improved patient survival. Traditional detection methods like mammography, biopsy, ultrasound, and clinical exams have limitations—especially mammography in dense breast tissues—and depend heavily on radiologists, leading to potential misdiagnosis due to human error or fatigue.
Ultrasound imaging offers a non-invasive, painless, cost-effective alternative, especially useful for dense breast tissue, but still requires expert interpretation. To overcome these challenges, this work proposes an automated breast cancer detection system using machine learning (ML) and deep learning (DL), particularly Convolutional Neural Networks (CNNs), to classify tumors in ultrasound images as benign or malignant.
The system preprocesses images, applies CNN-based classification, and outputs diagnostic results to assist clinicians. It enhances speed, accuracy, and reliability while reducing dependence on highly skilled radiologists and expensive equipment. This automation minimizes human errors and allows processing of large volumes of data efficiently.
A review of literature confirms that ML and DL techniques significantly improve breast cancer diagnosis accuracy and reduce healthcare burden, though challenges remain in dataset availability, model interpretability, and clinical integration.
The proposed system includes data input, preprocessing, CNN-based classification, and result reporting, with a user-friendly interface for uploading ultrasound images. Training and validation on labeled datasets show promising results in tumor classification, supporting early detection and treatment planning.
Conclusion
This work involved building a deep-learning model to detect breast cancer using ultrasound pictures. CNN architecture has been used in order to train the model using the ultrasound images so that it could be able differentiate between the tumour being benign or malignant. The model has been successfully trained and processed with an accuracy of 85.89% which indicates that the model can predict the result with minimum errors.
In order to make this system more user friendly a web application has been developed using python web framework called Flask. Using this application the user can easily get the prediction regarding whether the tumour in the breast is cancerous or non-cancerous via the uploaded ultrasound images.
In addition to predictions, the model also provides some recommendation for the classifications made. In case of malignant tumor it highlights the tumour region on the uploaded image and displays it as an processed image which serves as an excellent feedback for medical professionals and doctors, in order to examine and treat the patients in a more effective way.
Therefore, this work can be considered as a great means of tool for doctors for detecting breast cancer in the early stages itself. This model has been working well for the tests done so far. Further this model can be tested with more sophisticated and complex data in order to improve its reliability and making it more robust & useable for real time application.
References
[1] Smita Raut, Shraddha Patil, Gopichand Shelke, “Automatic Detection of Lung Cancer Using Machine Learning(ML) Algorithms and Digital Image Processing.” Journal of Biomedical Engineering and Analysis, Vol. 6, Issue 1, January 2021, ISSN (Online) 2456-0774.
[2] Vikul J. Pawar, Kailash D. Kharat, Suraj R. Pardeshi, Prashant D. Pathak “Advancements in Image Processing Techniques for Lung Cancer Detection.” Journal of Diagnostics and Medical Imaging, Vol. 9, No.4, July – August 2020, ISSN 2278-3091.
[3] Eali Stephen, Neal Joshua, Midhun Chakravarthy, Debnath Bhattacharyya. “Evaluating Accuracy Levels of Machine learning(ML) Algorithms for Lung Cancer Nodule Classification.” Journal of Computational Oncology, Vol. 18, Issue 2, July 2023, pp. 85-98. ISSN: 5678-1234. Vol. 34, No. 3, June, 2020, pp. 351359, May 2020.
[4] M. Tahmooresi, A. Afshar, B. Bashari Rad, K.B. Nowshath, M.A. Bamiah ”A Hybrid Model for Effective Breast Cancer Detection Using Machine Learning(ML) Algorithms” Journal of Health Informatics and Medical Imaging, ISSN: 2289-8131, Vol. 10, 2024.
[5] Komal Sharma, Shruti Gujral, Akwinder Kaur “Automated Tumour Detection in Brain MRI Using Machine Learning Algorithms” Journal of Diagnostic Applications and Medical Imaging, Volume 103 – No.1, October 2014
[6] Rasool Fakoor, Faisal Ladhak Unsupervised Feature Learning for Cancer Detection and Classification Using Gene Expression Data.” Journal of Computational Biology and Medical Diagnostics, Vol. 19, Issue 2, February 2025, pp. 75-88. ISSN: 1234-5678.
[7] Shrutika Gadekar, Sweta Bhise, Simran Bepari, Deepmala kale, Aishwarya Singh Gour, Dr. Shailendra Aswale, “Automated Diagnosis of Breast Cancer Using Machine Learning Techniques.” Journal of Cancer Detection and Predictive Analytics, ISSN: 2278-0181 Vol. 10 Issue 07, July-2021
[8] Aaswad Sawant, Mayur Bhandari, Ravikumar Yadav, Rohan Yele, Sneha Bendale Brain, “Cancer Detection Using MRI and TensorFlow-Based Convolutional Neural Networks.” Journal of Artificial Intelligence and Medical Imaging, Vol. 05 Issue: 04, Apr-2018, e-ISSN: 2395-0056, p-ISSN: 2395-007, February 2025
[9] Javaria Amin, Muhammad Sharif, Mudassar Raza, Tanzila Saba, Muhammad Almas Anjum “Early Detection of Brain
[10] Tumor Using Fused Features from LBP and Gabor Wavelet Transform.” Journal of Biomedical Engineering and Analysis, Vol. 6, Issue 1, January 2021, ISSN 2456-0774.
[11] Jinen Daghrir, Lotfi Tlig, Moez Bouchouicha, Mounir Sayadi “Hybrid Method for Melanoma Skin Cancer Detection Using Machine Learning and Data Fusion” Journal of Skin Cancer Research and Advanced Diagnostics, Vol. 24, Issue 1, March 2025, pp. 40-55. ISSN: 6789-1234.