Optimal Parameters for Modified Butterfly Interpolation Scheme Inspired Configurations as Hole-filling Method in 3D Volume Reconstruction

Three-dimensional volume reconstruction can help physicians to understand the patients' inner anatomy better in a 3D perspective. However, the pixel nearest neighbor method volume reconstruction tended to blur the reconstruction result. This study aims to present the design and configuration of improved modified butterfly interpolation scheme as the hole-filling method in the pixel nearest neighbor. The optimal parameters are also investigated to obtain a high-quality volume reconstruction result using a modified butterfly interpolation scheme. The experiment is designed and conducted on the ultrasound and CT scan data sets, where the artificial holes are created in these ground truth data sets. The proposed method is then used to reconstruct the missing voxels, and its capability is evaluated quantitatively. The results showed that the optimal butterfly interpolation parameter w is equal to 1/32 and 0 in the 3D reconstruction of ultrasound and CT scan, respectively. Meanwhile, the optimal maximum neighbor interpolation distance n is equal to 2 and 0 in the 3D reconstruction of ultrasound and CT scans, respectively.