Unsupervised Infrared Small-Object-Detection Approach of Spatial–Temporal Patch Tensor and Object Selection

Round 1

Reviewer 1 Report

Comments are addressed. 

Reviewer 2 Report

Dear Authors,

You have replied all my concerns. The paper can be accepted.

All the best.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Dear Author,

It is seen that your proposed approach is properly performed for small object detection. I just have 3 comments below:

1. Line 41-42, what do you mean by “infrared objects”? Is there any material that is called infrared? Please check and revise the whole text.
2. Line 54, please just use “ISTS”.
3. Line 101, “with the” should be removed.

All the best.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The authors have proposed a methodology to detect small objects in thermal images. The paper is hard to read and follow. The authors have made statements or provided alternatives without proper justification or references. The datasets are not properly introduced and cannot be considered diverse enough to assess the efficiency of the proposed work. The size of the object of interest has not been discussed at any point, which seems to be a very important factor. Additionally, the effect of environmental condition, such as ambient temperature, on the ability of the proposed method to detect a UAV is also not discussed. Another issue is lack of comparison with deep learning models for object detection such as CNN, R-CNN, and FCN. Some other minor comments are listed below:

Page 2, paragraph 2: some information is repetitive;

-Page 3 lines 99-100 these parameters must be defined. At the end of this page, it is good to see the limitations associated with the current methodologies visually.

Equation 2 needs more elaboration. What are the eigenvalues of an image? All parameters must be defined.

-what is justification for equation 3?

-what size of windows did the authors ended up using after ADMM is solved.

-what was the ground truth? More information about the dataset is required. Using only five images is not very convincing.

Table 1: what is SCR and why is its value missing for sequence 1? How different the objects are in terms of size (in terms of pixel). Additionally, are all thermal images similar in terms of size and resolution, and distance to the target?  

The link for dataset does not work. Why were this specific frame chosen? Why no use all the images. In figure 8, all images have rectangles.

-in Equation 20-22;

In Table 2, are these parameters selected based on the original paper's recommendation, or have the authors tuned them on their dataset?

 

 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

1. Image processing methods in recent years mainly use artificial intelligence, including artificial neural networks, so-called deep learning. One of the most common applications of object detection methods in images are small object detection. This article presents one of the problems related to the approach of spatial-temporal patch tensor and object selection.
2. General remarks
   1. Too many abbreviations make it difficult to follow the content of the article. Each abbreviation should be expanded the first time it appears. Not all readers need to know all abbreviations. Especially in the abstract of the article. What mean line 21 “F1 measure”?. “F1 measure” is used many times in the body of the article but not explained,
   2. Please use the language of a scientific research report without personal references: line 123 “Our”, line 485 ‘us’, lines 7, 128, 134, “we propose”, line 173 “we introduce”, line 183, 191 “we construct”, line 196 “we perform”, line 240 “we cannot”, line 318 “we choose”, line 327 ‘we denote’ and many others in whole article.
   3. The bibliography includes mainly domestic authors. Please do an in-depth literature analysis of the topic including works by authors from other continents.
   4. However the article is very well written should be carefully edited. Some remarks included below.
3. Specific remarks
   1. There should always be a space before the parenthesis and the rest of text (line 52, 53, 138,139).
   2. What means “Wher” – line 220, 228; “Wh” – line 237, 275, 281; “Whe” – line 247, 262?
   3. Line 354 – unneeded “Where”
   4. Line 447, 452- what means “frame 165”? And similar line 459,466 “frame 28 of sequence 16” and others in description of Fig. 12?
   5. Only initials are used in the chapter outlining the authors’ contributions.

Author Response

Please see the attachment

Author Response File: Author Response.docx