Next Article in Journal
Predicting Spatiotemporal Demand of Dockless E-Scooter Sharing Services with a Masked Fully Convolutional Network
Previous Article in Journal
Do Migrant and Native Robbers Target Different Places?
Previous Article in Special Issue
A Visual SLAM Robust against Dynamic Objects Based on Hybrid Semantic-Geometry Information
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
IJGI
Volume 10
Issue 11
10.3390/ijgi10110772
Review Report
ijgi-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
Article
Peer-Review Record

EnvSLAM: Combining SLAM Systems and Neural Networks to Improve the Environment Fusion in AR Applications

ISPRS Int. J. Geo-Inf. 2021, 10(11), 772; https://doi.org/10.3390/ijgi10110772
by Giulia Marchesi 1, Christian Eichhorn 1,*, David A. Plecher 1, Yuta Itoh 2 and Gudrun Klinker 1
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Reviewer 3: Anonymous
ISPRS Int. J. Geo-Inf. 2021, 10(11), 772; https://doi.org/10.3390/ijgi10110772
Submission received: 1 September 2021 / Revised: 25 October 2021 / Accepted: 6 November 2021 / Published: 12 November 2021
(This article belongs to the Special Issue Deep Learning for Simultaneous Localization and Mapping (SLAM))

Round 1

Reviewer 1 Report

In the case of the experiment, a high-performance desktop PC was used,  but in the abstract,  you said that your target device is smartphone and it seems your testing does not match with this.

It is common to write captions for tables and figures, and to give detailed explanations in the main text. However, most tables and figures do not have captions, but have explanations. Details of tables and figures should be explained in the text.

In general, AR using GPS assumes a mobile environment and a development environment for a smartphone rather than a PC. However, in the case of segmentation with NN that was presented in this text, high-performance PC and high power are required. So, it is questionable whether this can be implemented with EnvSLAM at Smartphone.

Author Response

Please see "Reviewer 1.pdf"

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear Authors,

Please find the attached file for my comments and please update the paper based on the comments. 

Best Regards.

Comments for author File: Comments.docx

Author Response

Please see "Reviewer 2.pdf".

Author Response File: Author Response.pdf

Reviewer 3 Report

This article is well written and present a SLAM system combined with a segmentation and semantic enhancement. The component of work(ORB-SLAM2 and Semantic SLAM) are open source codes and the uniqueness is adaption to mobile devices and integration of GPS signal based semantics extracted from Open Street Map. This work is a nice demonstration that AI can be run on edge devices and it can be used in multiple applications for real time detection of objects. It is well documented the work and it may be useful for other people to reproduce the Natural Phone results. The method is computationally intensive but can be carried out on a phone.

For results in Table 4, PA is PC PA I would have expected PC PA to be larger than PA.Can the author clarify why PA is larger than PC PA?

Author Response

Please see "Reviewer 3.pdf"

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Dear Authors, 

Thank you for addressing all my comments and I don't have any further concerns. The paper is accepted from my side. 

Best Regards 

Back to TopTop