Data availability

[nicill]

Dear authors,

Thank you very much for sharing your code.

Would it be possible to access the data used for the paper?

[CCRS-UAVLiCNN]

Hello,

Hope you are doing well. Which data are you looking for? The UAV orthomosaics?

Best, Galen

From: nicill @.> Sent: Wednesday, May 17, 2023 12:11 AM To: CCRS-UAVLiCNN/UAV_LiCNN @.> Cc: Subscribed @.***> Subject: [CCRS-UAVLiCNN/UAV_LiCNN] Data availability (Issue #2)

Caution - email originated from outside of NRCan. Read the warning below / Attention- Ce courriel provient de l'extérieur des RNCan. Voir la mise en garde ci-dessous

Dear authors,

Thank you very much for sharing your code.

Would it be possible to access the data used for the paper?

- Reply to this email directly, view it on GitHubhttps://github.com/CCRS-UAVLiCNN/UAV_LiCNN/issues/2, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AU27RKYYCJZ3E5CA2CLUFXLXGRFV5ANCNFSM6AAAAAAYEQC3ME. You are receiving this because you are subscribed to this thread.Message ID: @.***> This email originated from outside of NRCan. Do not click links or open attachments unless you recognize the sender and believe the content is safe. For more information, please visit How to Identify Phishinghttps://intranet.nrcan.gc.ca/services-policies/dont-get-scammed-cyber-security-101 emails on the NRCan Intranet. Ce courriel provient de l'extérieur des RNCan. Ne cliquez pas sur les liens et n'ouvrez pas les pièces jointes, à moins de connaître l'expéditeur et croire que le contenu est sécuritaire. Pour de plus amples renseignements, veuillez consulter Comment identifier des courriels d'hameçonnageshttps://intranet.nrcan.gc.ca/services-policies/dont-get-scammed-cyber-security-101 dans l'intranet des RNCan.

[nicill]

We are writing a review on deep learning techniques for the processing of UAV-acquired images. At this moment we are trying to assess what papers of those reviewed have data that could be accessed by future reserachers.

As part of the review we intend to run some simple experiments aiming at, for example, evaluating the performance of commonly used deep learning networks with different classification and segmentation problems. We are interested in any annotated data that can help us shed light onto the problems studied in the different papers and on how close to practical use the algorithms described are.

In the case of this paper, we would be most interested in the manually corrected masks mentioned in section 2.1 along with the corresponding images (if possible including the resaampled versions mentioned in section 2.1.2.)

Orthomosaics such as the one depicted in figure 5, are also very interesting, but if I read the paper correctly, annotations are not available for these so we would be unable to use them to experiment (we are not Lichen specialists and we do not have access to the site, so we are not able to do new annotations). Similarly, the data in section 3.3 is also very interesting because it speaks about a direct practical application, so if the 77 images mentioned along with the "Truthed values" depicted in figure 8 are available that would also be very interesting for us.

I realize this issue thread may not be the most practical way of communicating. Maybe we could continue this disussion via email and then I could send some more details about myself and my team. Would it be ok if I sent an email to the email listed for the corresponding author in the paper?

[CCRS-UAVLiCNN]

Hello,

My name is Galen Richardson and I am the first author on the paper, this is an ok method for communication, but it would be great if in your next email you could provide your name, affiliation, and email address.

This paper sounds very interesting, and I can provide the annotated data from 2.1.2 which I used to train my model. I feel like I have learned a lot since I wrote this paper, so you could get better results than what I published. As for the 77 annotated ground photos and cleaned masked (2.1), I will have to ask a few colleagues, but I think that it will be unlikely we can share those. There are not fully our IP.

If you wanted to simplify the data from section 2.1.2 for your paper, you could look into merging classes 2 and 3 to create binary images (No lichen, and pixels that contain lichen)

As for the orthomosaics, they are unavailable right now but I have a colleague who is looking to publish them. I can notify you once they are published but it will most likely still be a few months - a year before we get all the data sharing agreements sorted.

Best, Galen

From: nicill @.> Sent: Thursday, May 18, 2023 7:01 PM To: CCRS-UAVLiCNN/UAV_LiCNN @.> Cc: CCRS-UAVLiCNN @.>; Comment @.> Subject: Re: [CCRS-UAVLiCNN/UAV_LiCNN] Data availability (Issue #2)

Caution - email originated from outside of NRCan. Read the warning below / Attention- Ce courriel provient de l'extérieur des RNCan. Voir la mise en garde ci-dessous

We are writing a review on deep learning techniques for the processing of UAV-acquired images. At this moment we are trying to assess what papers of those reviewed have data that could be accessed by future reserachers.

As part of the review we intend to run some simple experiments aiming at, for example, evaluating the performance of commonly used deep learning networks with different classification and segmentation problems. We are interested in any annotated data that can help us shed light onto the problems studied in the different papers and on how close to practical use the algorithms described are.

In the case of this paper, we would be most interested in the manually corrected masks mentioned in section 2.1 along with the corresponding images (if possible including the resaampled versions mentioned in section 2.1.2.)

Orthomosaics such as the one depicted in figure 5, are also very interesting, but if I read the paper correctly, annotations are not available for these so we would be unable to use them to experiment (we are not Lichen specialists and we do not have access to the site, so we are not able to do new annotations). Similarly, the data in section 3.3 is also very interesting because it speaks about a direct practical application, so if the 77 images mentioned along with the "Truthed values" depicted in figure 8 are available that would also be very interesting for us.

I realize this issue thread may not be the most practical way of communicating. Maybe we could continue this disussion via email and then I could send some more details about myself and my team. Would it be ok if I sent an email to the email listed for the corresponding author in the paper?

- Reply to this email directly, view it on GitHubhttps://github.com/CCRS-UAVLiCNN/UAV_LiCNN/issues/2#issuecomment-1553760572, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AU27RK2ZLWQ5ZWLGGIMK2UDXG2S3ZANCNFSM6AAAAAAYEQC3ME. You are receiving this because you commented.Message ID: @.***> This email originated from outside of NRCan. Do not click links or open attachments unless you recognize the sender and believe the content is safe. For more information, please visit How to Identify Phishinghttps://intranet.nrcan.gc.ca/services-policies/dont-get-scammed-cyber-security-101 emails on the NRCan Intranet. Ce courriel provient de l'extérieur des RNCan. Ne cliquez pas sur les liens et n'ouvrez pas les pièces jointes, à moins de connaître l'expéditeur et croire que le contenu est sécuritaire. Pour de plus amples renseignements, veuillez consulter Comment identifier des courriels d'hameçonnageshttps://intranet.nrcan.gc.ca/services-policies/dont-get-scammed-cyber-security-101 dans l'intranet des RNCan.

[nicill]

I have sent you an email to the address that appears in the paper, closing the issue, thanks for the attention.