Skip to content

Panpan-Chen/Attention-Block-U-net

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

15 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Attention Block U-Net

Improved Photoacoustic Imaging of Numerical Bone Model Based on Attention Block U-Net Deep Learning Network

Source code for ''Improved Photoacoustic Imaging of Numerical Bone Model Based on Attention Block U-Net Deep Learning Network''.

Models

We designed an Attention Block U-Net (AB U-Net) Network from the standard U-Net by integrating the attention blocks in the feature extraction part, aiming to be more adaptive for imaging bone samples with complex structure.

image

The attention blocks originated from Convolutional block attention module (CBAM).

U-Net and Attention U-Net models are also contained in this repository.

Train

python main.py

Test

Run test.py

Visualization

The curves of BCE loss, PSNR and SSIM with iterations are realized by TensorBoard.

# runs tensorboard --logdir runs

Results

Visual comparison of the performance in three examples based on Time Reversal and AB U-Net

AB U-Net successfully removes artifacts and restores the high-frequency information, such as the micro-structure of the trabecular bone. Compared with Time Reversal method, the CNN-based network provides significant improvement in PSNR and SSIM, i.e., SSIM of sample 1 increases from 0.62 to 0.88, indicating an accurate modeling of the initial pressure.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages