MLGNet: Multi-Task Learning Network with Attention-Guided Mechanism for Segmenting Agricultural Fields

Round 1

Reviewer 1 Report

The manuscript discusses the implementation of a multi-task learning network with an attention-guided mechanism for segmenting agricultural fields using high-resolution remote sensing images. The network incorporates an attention-guided fusion module to capture complementary information and considers both edge detection and semantic segmentation tasks. This work demonstrates the potential of the proposed method for precise agricultural development, particularly in dealing with fields that have blurred edges.

 

The research looks valuable and completed. I would improve it in the way, 

 

1. Taking time series in observational data

2. Increasing the number of studied sites.

3. Providing more details on potential applications. 

check the articles if it is possible. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

great work.

It might be better if the authors do not use the word "we". 

For example, in the abstract, the sentence "we introduce a multi-task learning network with attention-guided mechanism for segmenting agricultural fields", can be rewritten as " a multi-task learning network with attention-guided mechanism was introduced for segmenting agricultural fields".

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

The authors proposed a multi-task learning network with attention-guided mechanism from segmenting agricultrual fields. One of the main issues is the quality of the figures, all the figures are with a very lower quality, I cann't read the figures correctly. Another  is the limited novelty of the proposed method. The authors combine different building blocks from deep learning community. Although the final results seem good, the authors should give more experiments on the  effectiveness of the proposed building blocks.  In this way,  the experiment will demonstrate the effectiveness of the proposed method.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

Article presents a multi-task learning network with an attention-guided mechanism for segmenting agricultural fields using high-resolution remote sensing images. The method considers both edge detection and semantic segmentation tasks and uses a connectivity perception theory to segment merged fields with blurred edges. The study area for experimentation was three cities in the Netherlands, and the results showed high accuracy and improvements in processed edges. The method has the potential for precision agriculture development.

 

Figure names are like paragraphs, my recommendation - should be shorter, and the text moved to the body of the article and referred to a name/figure in the text. For example Figure 2.

 

Line 583-539: These results indicate that your network framework and multi-task learning method are very effective.

If you are referring to a network framework that the authors developed or proposed, it would be more appropriate to use "our network framework" or "the proposed network framework" instead of "your network framework." This helps to clarify that the network framework is being attributed to the author or the research team.

For example:

"These results indicate that our network framework and multi-task learning method are very effective."

or

"These results indicate that the proposed network framework and multi-task learning method are very effective."

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

I have no other comments.