EU^2net: A Parameter Efficient Ensembled U^2N et Model with Attention-aided Triple Feature Fusion for Tumor Segmentation in Breast Ultrasound Images

This is the official implementation of "EU^2net: A Parameter Efficient Ensembled U^2N et Model with Attention-aided Triple Feature Fusion for Tumor Segmentation in Breast Ultrasound Images" (IEEE TIM 2024).

Overall workflow:

A2TF2 block:

FSiAM:

SE-CAM:

Training and validation curves:

Saved model: https://drive.google.com/drive/folders/12fxzkx0kUjuo-P6lSnwEKDZwYFprT59V?usp=drive_link

Results

Authors 🤓

Ayush Roy
Payel Pramanik
Ram Sarkar

Citation 🤔

Please do cite our paper in case you find it useful for your research.

@article{roy2024eu,
  title={EU 2--Net: A Parameter Efficient Ensemble Model with Attention-aided Triple Feature Fusion for Tumor Segmentation in Breast Ultrasound Images},
  author={Roy, Ayush and Pramanik, Payel and Sarkar, Ram},
  journal={IEEE Transactions on Instrumentation and Measurement},
  year={2024},
  publisher={IEEE}
}