Creatis- Univesity of Lyon || NTNU || Technical University of Denmark


Evaluation framework
Modifying DAS beamformer
  • Open the file +beamformer/das.m in the Matlab Editor
  • Save the file with other name, for instance das_2.m, but in the same folder +beamformer
  • Change the name of the function accordingly, for instance das_2.
  • Feel free to change the description, author names, and date.
  • The first code block loads the dataset and the scan from the path provided to the function handle. No need to change this block.
  • In case of RF data, the next block computes the Hilbert transform of the dataset. In this way the same pixel-based beamforming algorithm can be applied directly. This step may not be ecessary for your implementation. In particular, if your algorithm works only with RF or IQ data, here you can check that the input data has the right format.
  • Here we define the angular and receive apodization used in DAS. In particular a 1.75 F-number dynamic expanding aperture with 25% Tukey window is used on receive. A flat angular apodization is used to combine the plane-waves. You can decide freely what apodization values are used in your algorithm.
  • The next block contains the beamforming loop. This is a simple, easy-to-understand implementation of DAS and hence is not optimized for speed. You can substitute this block with you own beamforming algorithm.
  • After beamforming the code reshapes the pixel matrix and computes the signal envelope. The pixel values produced by your algorithm should be positive real number.
  • The last block saves the image to disk in hdf5 format. Compulsory inputs are the envelope of the beamformed data, the scan, and number of compounded waves. You should include the algorithm name as well as your names and institution. Optionally, if you used a conventional transmit or angular apodization, you can insert that data in the related fields.

Testing your beamformer

Assuming you followed the previous steps:
  • Open the file examples/script_generate_image.m in Matlab Editor
  • Modify the reconstruct block changing the name of the beamformer for the one you chose in the previous section, for instance das_2
  • Run the script and check that the image is displayed correctly
  • You can check the metrics of your method just by running the script examples/script_evaluate_image.m