Deep learning models for breast cancer detection benchmarked for performance and emissions

A new paper benchmarks seven deep learning models for breast cancer detection, evaluating their performance and environmental impact. The study found that while EfficientNet and ResNet offer strong accuracy, they also produce higher CO2 emissions. Transformer models like DeiT-Tiny, ViT, and Swin Transformer showed competitive results, with DeiT-Tiny offering a good balance of accuracy and energy efficiency on one dataset, and ViT and Swin excelling on another. The research emphasizes the need to consider performance, emissions, and dataset specifics when choosing models for medical applications. AI

IMPACT Highlights the trade-offs between model performance, energy consumption, and dataset characteristics in medical AI applications.

RANK_REASON Academic paper evaluating deep learning models. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• BreakHis 400X
• Breast Ultrasound Dataset
• Data Efficient Image Transformers
• DeiT-Tiny
• DenseNet121
• EfficientNet
• residual neural network
• Swin Transformer
• Vít

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →