Deep Learning-Based Real-Time Discriminate Correlation Analysis for Breast Cancer Detection.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Breast cancer is the most common cancer in women, necessitating accurate diagnostic tools.
• Current breast mass recognition models often suffer from overfitting due to a scarcity of medical image samples.
• Effective clinical diagnosis relies on robust and accurate image recognition models.

Purpose of the Study:

• To propose a breast mass recognition model that integrates deep pathological information mining to address data scarcity and overfitting.
• To enhance feature optimization and data augmentation strategies for improved model performance.
• To develop an efficient model for accurate mammography image classification.

Main Methods:

• A sample selection strategy was employed for high-quality sample screening across diverse mammography datasets.
• The multiview effective region gene optimization (MvERGS) algorithm was utilized to refine image features, enhance discrimination, and compress dimensions.
• Deep pathological information was mined through cross-modal correlation analysis (DCA) for accurate lesion area description.

Main Results:

• The proposed breast mass recognition model demonstrated superior accuracy and AUC compared to mainstream baselines.
• The model effectively alleviated the overfitting problem commonly associated with limited medical image samples.
• Feature refinement and deep pathological information mining led to accurate breast mass lesion area description.

Conclusions:

• The integrated approach of data enhancement, feature optimization, and deep pathological information mining significantly improves breast mass recognition.
• The developed model offers a promising solution for accurate and reliable clinical diagnosis of breast cancer.
• Addressing data scarcity through advanced algorithms is crucial for developing robust medical image recognition systems.