A Benchmark for the Evaluation of Corner Detectors
Sanjay Ghosh
Round 1
Reviewer 1 Report
- Before publication, in order to improve the paper, the authors to take into consideration all of the following major remarks to improve the quality of the presentation of their work. Besides, please review the grammar and adjust the verb tenses.
- 1. Literature is appropriate, however more references could be used to improve the introduction section.
- 2. There are few references in the past three years, and supplementary experiments are needed.
- 3. Results should be commented with more details, besides and discussion section should be cited some relevant references.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The paper introduces a dataset for corner detection containing a set of urban images, called that Urban-Corner dataset. Having such a benchmark dataset with labelled ground-truth corners and unified metrics would be very useful to the image processing community to evaluate corner detection performance of respective methods.
However the major limitation is the number of images - 21 is just very few.
The write-up of the paper may be improved. A quick summary of the at least some of the SOTA methods could attract more readers' attention.
I would encourage the authors to discussion on kernel filtering based methods [R1, R2] for corner/edge detection.
[R1] "Fast scale-adaptive bilateral texture smoothing,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 30, no. 7, 2020.
[R2] “On fast bilateral filtering using Fourier kernels,” IEEE Signal Processing Letters, vol. 23, no. 5, pp. 570-574, 2016.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The revised version of the paper may be accepted for publication.
