"Hybrid Model for Breast Cancer Detection Through Mammographic Images Using Vision Transformer Architectures and Machine Learning Techniques"

Abstract

"Breast cancer is one of the leading causes of death worldwide, and many cases are detected at advanced stages due to barriers to accessing healthcare services and timely diagnoses. In response to this problem, a robust hybrid model is proposed to detect this disease through mammographic images, integrating the capabilities of vision transformers with machine learning techniques. A five-phase methodology was applied: Obtaining the dataset, Preprocessing, Implementing hybrid models based on Deep Learning (Swin Transformer, MaxViT, BEiT, DeiT, and ConViT Small) and Machine Learning (XGBoost, MLP, and LightGBM), Model evaluation, and Prediction. The best results were obtained with the BEiT?+?MLP hybrid model with the following metrics: accuracy, precision, recall, and F1-score, whose results were 95.12%, 96.05%, 96.51%, and 96.27%, respectively. In conclusion, the results show that the application of hybrid models is highly effective in detecting cancer through mammographic images and based on this, making timely diagnoses and treatments. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2026."