Content of review 1, reviewed on December 15, 2018
1- You have to write UAV, abbreviation of what (unmanned aerial vehicle) at Abstract. please define all other abbreviations used in your manuscript (CRF (Conditional Random Field), HMM, PGT,...etc). 2- Page 1, 2nd paragraph, 1st line, could you rewrite it to make it more understandable? (brings about immense !!). 3- I think you have to say image annotation instead of image annotating, please change it. 4- Could you minimize the title of Figure 1, move this explanation to Introduction section? 5- Page 4 of 26, last line of the 2nd paragraph, has to be removed. 6- Page 5 of 26, Subsection 3.1, 3rd paragraph, why you have believed the transferred annotations have 80% accuracy, is this fact or you have depended on some References? 7- Page 6 of 26, 5th Line, explain why n is set to 2, n represents what? 8- Page 10 of 26, subsection 4.4.2. 5th Line, could you explain in more details why (each pixel in the UAV image corresponds to multiple 3D points which may carry different labels). 9- Page 12 of 26, 1st paragraph, c) illustrates the predictions deploying automatically generated, while in the titlle of Figure 9, c) predictions using manually annotated, and the same error for d, please check it. 10- Page 12 of 26, could you write some words to explain the meaning of Intersection over Union (IoU). 11- Page 16 of 26, subsection 5.3, 1st line, give the reference for this statement (manual annotations is not always correct). 12- Title of Figure 16, 2nd line, (bottom row) not bottom column. I think (c) has to be changed between 1st row and second row? because OSM has position accuracy while inferred building annotation is not accurate, check it. 13- Support conclusion section with some numerical results. 14- References title is not written.
Source
© 2018 the Reviewer.
References
Zhuo, X., Fraundorfer, F., Kurz, F., Reinartz, P. 2019. Automatic Annotation of Airborne Images by Label Propagation Based on a Bayesian-CRF Model. Remote Sensing.