Improving Multi-View Camera Calibration Using Precise Location of Sphere Center Projection
, Javier Perez-Soler 1, Juan-Carlos Perez-Cortes 1,2,* and Jose-Luis Guardiola 1
Round 1
Reviewer 1 Report
The paper is interesting, it only requires minor corrections and improvements.
- Authors did not take into account the influence of the lens optical aberrations on the recorded image distortion. It would be good to add a comment on this topic.
- In some places, the narrative of the paper is conducted in the first person instead of the third. For example "we did it" and instead this it should be use the form "it has been done".
- There are many editorial errors and mistakes like no space between the value and the unit or mistake in caption under the Figure 10. Please check the paper on this matter.
- There are only nine references and two of them are self-citations. The proportion between self-citations and all references should be different.
Author Response
The paper is interesting, it only requires minor corrections and improvements.
- Authors did not take into account the influence of the lens optical aberrations on the recorded image distortion. It would be good to add a comment on this topic.
RESPONSE:
This distortion is modelled with the camera intrinsic parameters and should be corrected over the image space (in pixel units) before to convert pixels units in millimeters. A comment is added in line 85.
- In some places, the narrative of the paper is conducted in the first person instead of the third. For example "we did it" and instead this it should be use the form "it has been done".
RESPONSE:
Reviewed.
- There are many editorial errors and mistakes like no space between the value and the unit or mistake in caption under the Figure 10. Please check the paper on this matter.
RESPONSE:
Corrected.
- There are only nine references and two of them are self-citations. The proportion between self-citations and all references should be different.
RESPONSE:
More references have been added and the introduction is extended.
Reviewer 2 Report
Dear Authors,
I have read the paper with great interest and have the following remark. In my opinion two major issues are missing (or at least are not emphasized enough): detailed review of the state-of-the-are and comparison with the state-of-the art. I would for example expect that several methods are compared on the same dataset, also with timing and computational complexity analysis (BTW. is time so important in the considered case?).
You could also provide some information about the whole calibration procedure. You often refer to your previous work [3], but for completeness at lest some information should be provided.
All in all, I suggest a major revision.
Some minor remarks:
Abstract
Consider improving (extending the abstract).
Section 1
Line 26 “And on the other” - should be connected to the previous sentence or other language construction used.
Section 2
Figure 1. is projected in the -> on the ?
Eq (5). W -> W^2 (right part of the equation). (the same in Alg. 1)
Section 3
Line 108 algorithm. em- (remove dot).
Reviewer
Author Response
I have read the paper with great interest and have the following remark. In my opinion two major issues are missing (or at least are not emphasized enough): detailed review of the state-of-the-are and comparison with the state-of-the art. I would for example expect that several methods are compared on the same dataset, also with timing and computational complexity analysis (BTW. is time so important in the considered case?).
RESPONSE:
A more detailed introduction is written to present other existing approaches.
A comparison between method is may be not interesting because all the methods propose a valid solution for the sphere position estimation or the sphere center projections correction. The differences lie in the data employed to do the correction or in the methods (exact, approximate, geometrical, ...).
Regarding the time complexity, if the correction is needed to calibrate a set of cameras, that is not very important because normally this is done off-line (once every time the system is reconfigured, repaired, ...) In our work it is mentioned only because an iterative method is proposed, then if it converges slowly could be a drawback in contexts were a fast computation is needed. But as stated in the paper it is not the case.
You could also provide some information about the whole calibration procedure. You often refer to your previous work [3], but for completeness at lest some information should be provided.
RESPONSE: A new figure is introduced to clarify the calibration procedure and a more detail on the procedure are added. If necessary, more description can be added.
All in all, I suggest a major revision.
Some minor remarks:
Abstract
Consider improving (extending the abstract).
RESPONSE: done.
Section 1
Line 26 “And on the other” - should be connected to the previous sentence or other language construction used.
RESPONSE: paragraph modified.
Section 2
Figure 1. is projected in the -> on the ?
Eq (5). W -> W^2 (right part of the equation). (the same in Alg. 1)
RESPONSE: fixed.
Section 3
Line 108 algorithm. em- (remove dot).
RESPONSE: fixed.
Reviewer 3 Report
Dear Authors,
I have read the manuscript with attention and interest. The article as a whole (composition of work, subject matter) made a positive impression on me and during reading, I received some remarks, comments, and proposed changes that I hope will serve to improve the manuscript.
The article is written well, and easy to read.
I consider the number of references and thus the introduction/conclusion to be a weak part. 9 sources are used, of which 2 are self-citation and 5 are more than 20-30 years old (which would not be problem if there are a sufficient number of other references). In the introduction, there are not mentioned any other works, which presents different approach to the problem. Subsequently, in conclusion, there is missing comparison/advantages/improvements of your method and results to the other works/approaches. Accordingly, the introduction, conclusion, and references section must be broadened.
Formal suggestions:
L170 is the same as L174
L171 is the same as L175
Figure 4 could be rendered after it is mentioned in the text.
Figure 9 could be rendered after it is mentioned in the text.
Figure 9 missing units for X and Y
Table 1 and 2 are not labeled according to the article template
missing References title
Author Response
I consider the number of references and thus the introduction/conclusion to be a weak part. 9 sources are used, of which 2 are self-citation and 5 are more than 20-30 years old (which would not be problem if there are a sufficient number of other references). In the introduction, there are not mentioned any other works, which presents different approach to the problem. Subsequently, in conclusion, there is missing comparison/advantages/improvements of your method and results to the other works/approaches. Accordingly, the introduction, conclusion, and references section must be broadened.
RESPONSE: The introduction has been extended giving more information and references over other related works. Similarly in the conclusions more details are included.
Formal suggestions:
L170 is the same as L174
L171 is the same as L175
RESPONSE: In our version no duplicated lines exists.
Figure 4 could be rendered after it is mentioned in the text.
RESPONSE: Done.
Figure 9 could be rendered after it is mentioned in the text.
RESPONSE: Done.
Figure 9 missing units for X and Y
RESPONSE: Done.
Table 1 and 2 are not labeled according to the article template
RESPONSE: Table and document format adapted.
missing References title
RESPONSE: Added.
Round 2
Reviewer 2 Report
Dear Authors,
Thanks for responses. I am quite satisfied. I would only slightly modify the abstract (details below) and conclusion (more details).
Comparison with SOTA - OK
Time complexity - OK
Introduction
- segmentation image -> image segmentation ?
- "Accuracy...." -> verb missing ? like "occur" ?
Abstract - for your own benefit (to attract readers) put some details - like what are the mentioned benefits.
Other issues - OK
Author Response
Thanks for responses. I am quite satisfied. I would only slightly modify the abstract (details below) and conclusion (more details).
RESPONSE: Done.
Introduction
segmentation image -> image segmentation ?
"Accuracy...." -> verb missing ? like "occur" ?
RESPONSE: Fixed.
Abstract - for your own benefit (to attract readers) put some details - like what are the mentioned benefits.
RESPONSE: Done.
Reviewer 3 Report
Dear Authors.
You have addressed all my remarks and comments.
Your article just need english correction: e.g. L28 29 analitical - analytical
Author Response
Your article just need english correction: e.g. L28 29 analitical - analytical
RESPONSE: Spell check and other grammatical revisions performed.
