Soft Tissue/Bone Decomposition of Conventional Chest Radiographs Using Nonparametric Image Priors

Algorithm 1

Hierarchically dense matching.

Input: Two image A (source image) and B (target image);

Dense local descriptors F_A of A and F_B of B.

Output: Corresponding field T.

1: Divide A to the non-overlapping cells of size w×w and obtain the BoW histograms for each cell i.

2: Divide B to overlapping cells of size w×w with step size s and obtain the BoW histograms of each cell j.

3: Determine the most similar cell centered at (x_B, y_B) in B for each cell c_A in A.

4: Set T(x, y)=(x_B, y_B) ((x, y)∈c_A) for each cell c_A in A and random permute of T.

5: Update T by field propagation as PatchMatch algorithm using the local descriptors F_A and F_B.

6: Update T by locally randomized search using the local descriptors F_A and F_B.

7: Repeat steps 5 and 6 a fixed number of times or until convergence.