Generating a synthetic two-dimensional mammogram
A method for generating a synthetic two-dimensional mammogram with enhanced contrast for structures of interest includes acquiring a three-dimensional digital breast tomosynthesis volume having a plurality of voxels. A three-dimensional relevance map that encodes for the voxels the relevance of the underlying structure for a diagnosis is generated. A synthetic two-dimensional mammogram is calculated based on the three-dimensional digital breast tomosynthesis volume and the three-dimensional relevance map.
1. A method for generating a synthetic two-dimensional mammogram with enhanced contrast for structures of interest, the method comprising:
acquiring, with a medical scanner, a three-dimensional digital breast tomosynthesis volume having a plurality of voxels;
generating a three-dimensional relevance map that encodes for the plurality of voxels a relevance of an underlying structure for a diagnosis; and
calculating a synthetic two-dimensional mammogram based on the three-dimensional digital breast tomosynthesis volume and the three-dimensional relevance map,
wherein calculating the synthetic two-dimensional mammogram is based on ray-casting in the generated three-dimensional relevance map and the acquired three-dimensional digital breast tomosynthesis volume.
2. The method of claim 1 , wherein the three-dimensional relevance map is generated based on a predefined generating model from the acquired three-dimensional digital breast tomosynthesis volume.
3. The method of claim 2 , wherein the predefined generating model is learned by analyzing a set of three-dimensional digital breast tomosynthesis volumes.
4. The method of claim 3 , further comprising computing features at the plurality of voxels of the three-dimensional digital breast tomosynthesis volumes to describe a local visual context.
5. The method of claim 1 , further comprising creating a plurality of channels from the acquired three-dimensional digital breast tomosynthesis volume.
6. The method of claim 5 , wherein the plurality of channels is created based on gradient magnitudes, gradient orientations, or Laplacians.
7. The method of claim 6 , further comprising determining relevance values of the three-dimensional relevance map, the determining of the relevance values comprising separately analyzing the plurality of channels.
8. The method of claim 1 , further comprising post-processing the generated three-dimensional relevance map to minimize the impact of noisy predictions.
9. The method of claim 8 , wherein the post-processing is performed by using a Gaussian filtering of the generated three-dimensional relevance map.
10. The method of claim 1 , further comprising adding a global constant weight to the generated three-dimensional relevance map.
11. The method of claim 1 , wherein calculating the synthetic two-dimensional mammogram is based on weighted averaging along determined rays.
12. The method of claim 1 , wherein calculating the synthetic two-dimensional mammogram is based on determining a maximum weighted value found along rays.
13. The method of claim 1 , further comprising displaying the synthetic two-dimensional mammogram on a screen.
14. A mammographic device for generating a synthetic two-dimensional mammogram with enhanced contrast for structures of interest, the mammographic device comprising:
an imaging device configured to acquire a three-dimensional digital breast tomosynthesis volume having a plurality of voxels; and
one or more processors configured to:
generate a three-dimensional relevance map that encodes for the plurality of voxels a relevance of an underlying structure for a diagnosis; and
calculate a synthetic two-dimensional mammogram based on the three-dimensional digital breast tomosynthesis volume and the three-dimensional relevance map,
wherein the calculation of the synthetic two-dimensional mammogram is based on ray-casting in the generated three-dimensional relevance map and the acquired three-dimensional digital breast tomosynthesis volume.
15. In a non-transitory computer-readable storage medium storing instructions executable by a computer to generate a synthetic two-dimensional mammogram with enhanced contrast for structures of interest, the instructions comprising:
acquiring a three-dimensional digital breast tomosynthesis volume having a plurality of voxels;
generating a three-dimensional relevance map that encodes for the plurality of voxels a relevance of an underlying structure for a diagnosis; and
calculating a synthetic two-dimensional mammogram based on the three-dimensional digital breast tomosynthesis volume and the three-dimensional relevance map,
wherein calculating the synthetic two-dimensional mammogram is based on ray-casting in the generated three-dimensional relevance map and the acquired three-dimensional digital breast tomosynthesis volume.