Reconstructed results are shown in
where the pink-colored region (at 3 o’clock) clearly indicates the location and size of the object (, , , , , , and ). As can be seen, the images formed are shown to be qualitatively correct in visual content for all the cases. To provide a more quantitative assessment of these images, the reconstructed optical property distributions are displayed along one transect through the center of reconstructed target—the one perpendicular to the line through the centers of both the target and background regions—compared with the exact values, in , , , , , , and .
provides the quantitative information about the center and the full-width-half-maximum (FWHM) of the object as well as the recovered average optical properties of the object estimated from a region-of-interest (ROI) determined by the center and the FWHM of each object.
Fig. 3 Reconstructed absorption coefficient images and profiles through the transect from PAT and DOT. (a) and (b): Case 1 from PAT; (c) and (d): Case 1 from DOT; (e) and (f): Case 2 from PAT; (g) and (h): Case 2 from DOT; (i) and (j): Case 3 from PAT; (k) and (more ...)
Reconstructed reduced scattering coefficient images and profiles through the transect from DOT. (a) and (b): Case 2, (c) and (d): Case 3.
Reconstructed Values of Absorption and Reduced Scattering Coefficients (mm−1) of the Target and Background and Target Location (Off-Center) and Size (mm) for the Phantom Experiments*
From the recovered absorption coefficient images, while we can see that the value of absorption coefficient is well reconstructed both from DOT and PAT data for all the cases, the location and size of target are much better recovered by PAT than that by DOT. For example, the relative errors of the recovered absorption coefficient by DOT are, 2%, 14%, and 4%, for cases 1-3, respectively, while the relative errors of the reconstructed absorption coefficient by PAT are, 16%, 1%, and 3%, for cases 1-3, respectively. For the DOT images, while the target is detected, we note that the reconstructed target position is slightly shifted relative to the exact target position. The PAT recovered target size is found to be 7.1mm, 5.3mm, and 6.1mm for the three cases, respectively, whereas the DOT recovered target size (11-14mm) is significantly overestimated compared to the exact size.From , we noticed that the recovered target size and absorption value are more accurate when the contrast is increased for both DOT and PAT. The PAT image showed some target distortion () when only absorption contrast existed. This could be due to the use of a small number of transducers and the lack of target scattering contribution to the PAT reconstruction through the optical fluence/photon density. This can be improved by adding more transducers to the PAT system.
We also noted that some artifacts appear in the DOT images especially for case 1 (). These artifacts generally are shown near the boundary source and detector positions where the measurement sensitivity is highest. However, such boundary artifacts are not seen in the PAT absorption images.
From , we see that the scattering images for both cases are recovered with high quality without boundary artifacts by DOT. In addition, we found that the recovered size of the target ranges from 5.9mm to 6.2mm, in good agreement with the actual object size of 6mm. The recovered off-center location of the target ranges from 9.4mm to 9.6mm, also in good agreement with the actual off-center location of 10mm. Finally, we see that the scattering coefficient values of target and background are well recovered quantitatively as seen from .