Limited-Angle Computer Tomography with Truncated Projection Artifacts Removal
Round 1
Reviewer 1 Report
In this study the new method for elimination of the truncated projection artifacts (TPA) from the digital breast tomosynthesis (DBT) mammography images is being presented. From an application point of view, this is very important issue - the question is how often it is encountered in a daily clinical practice.
The remarks to the article are listed below.
The captions of Figures 2 and 9 are not described sufficiently.
There is something wrong with numbering of the figures. In lines 123, 126, and 135 there is reference to Figure 4 which is not present at all. I guess it should refer to Fig. 3 and the rest of figures should be renumbered.
Line 132 - the referenced number of the equation should be in the brackets.
The description of Figure 6 is missing.
Line 166 - which diagram do you mean? This presented on the supposed Fig. 6?
In Algorithm 2 - in the last step of the algorithm - step 6 - how can we know that the artifact is already eliminated (especially when we deal with real data)?
Please provide in Section 3.1 information about the software/programming language by which the simulations were carried out.
Line 246 - correct the references.
In Subsection 3.3 you calculate SSIM index for both SART and your algorithm. In the Result section simulations were also performed for the diffusion-based method - why SSIM index was not calculated also for this method?
It would be worth to mention what percentage of breast volume was affected by the TPA for both simulated and real data. I understand that in reality it strongly dependent on breast size and positioning during examination but it could give the reader insight on the scale of the problem.
Author Response
As attached.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The submitted manuscript describes a method for artifact removal. It might be of interest for the readers of Applied Sciences. Below are my comments that in my point of view may clarify certain points for the reader and improve the submitted manuscript
Comment 1 Line 8-10 Should the statement above the Abstract section be there? You may consider its removal.
Comment 2 Line 28 DBT is more an established than an emerging technique. The statement may was true several years ago, but in the present is not very accurate. Please change the word emerging.
Comment 3. Novelty. A clear statement of the manuscript novelty should exist. For instance a paper of some of the authors (Enxiang Shen, Yuxin Wang, Jie Yuan, Truncated Projection Artifacts Removal in Digital Breast Tomosynthesis, https://ieeexplore.ieee.org/abstract/document/9882423) is published as an IEEE conference proceeding. The methodology seems very similar, the phantom is different though. Please insert this reference in the reference list and also insert a thorough discussion in how the submitted work has novelty with regards to the proceeding.
Comment 4. Lines 47-53. The text is correct however, you should add for clarification that the extend of TPA can depend upon the sweep-angle, the breast size, the detector area and the acquisition pattern (continues and step and shoot).
Comment 5. Figures 1. Please take care of the figures numbering. Figure 1 is designated twice (page 2 and page 3). Also figure 4 is missing. Please correct figure numbering and where necessary their mention in the text.
Comment 6. Equation 1. This is a crucial equation and some extra explaining is needed. I propose to demonstrate graphically like in the old Figure 1b (probably Figure2b) the positions of t, R, c etc in a geometric figure, as well as the xyz coordinates.
Comment 7. Lines 97-112 Your xr(k+1,c) is the linear attenuation coefficient as calculated and updated from the projection method. Your initial data however are the pixel values of the detector, the angle and the breast thickness. The pixel value depends upon the response curve (mGy vs AirKerma) of your detector as well as the X-ray energy spectrum. I suppose you consider average values and the data will change depending upon the irradiation conditions in any case, but you should make a comment about it.
Comment 8, page 7 on top. There is a figure in the top of the page describing the steps, but there is no figure legend and number, at least at my pdf. Please check and correct if necessary. Please also see my comment 4.
Comment 9. The use of the gradient may demonstrate high signal changes, due to the TPA, but locally a microcalcification if exist may affect the method. Please comment. Also, use a reference for the technique.
Comment 10. At the descriptions of the steps and their graphical representation it is not very clear what p and q are. Do they correspond to the missing area (have area units?) or they are the distances between the virtual panel and the edge of the truncated part for each scan line? Please clarify.
Comment 11. lines 158-166. Did the authors compare there reconstructed data, or this is a point taken from literature. Please either insert some reconstructed data having the behaviour of lines 158-166, or use an appropiate reference.
Comment 12. lines 169-175. How did the authors include tumor, ligaments and other objects? Please state there dimensions and their signal values with respect to the background.
Comment 13. Lines 176-177, please show in Figure 2 or 1, where are the 50 mm with respect to the 60mm of the breast.
Comment 14. Figure 7, page 8. The reconstructed dimensions seems streched. Please calculated the distorion (dimension differences) between the original image and the reconstructed image. Is the effect of the presented algorithm in distortion more or less intense than the other algorithms? Please add text accordingly.
Comment 15.Figure 7, legend. I believe that "d" in the end should be "f". Please check and correct if needed.
Comment 16.Lines 244-245. The text writes algorithm25,26. If 25 and 26 are the reference numbers please correct the writting accordingly. In my opinion, reference 26 is useful for the manuscript but it is misplaced. It does not provide explicit information about SSIM.
Comment 17. I was not able to find any mention of references 11, 18, 21-24 in the manuscript. Please check again and if necessary correct.
with kind regards
The reviewer
Author Response
As attached
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The authors of the manuscript address my review comments in the revised document as well as in the answering letter. The clarity of the manuscript is increased.
In my revised version however, the figures are still not numbered correctly. Perhaps it was a missing point beween changes, but it should be fixed
So
in lines 60-61 there is Figure1
in lines 91-94 there is again Figure1
Figure 4 although referenced in lines 123-140 is missing (as a name)
Please check again and correct all the figure numbering and their corresponding reference in the text.
with kind regards
the reviewer
Author Response
We have checked all figure indexes.
