Automated Identification of Morphological Characteristics of Three Thunnus Species Based on Different Machine Learning Algorithms
, Qi He 5,*, Weiguo Qian 6,* and Leilei Zou 7
Round 1
Reviewer 1 Report
Your paper explores the use of ML/AI to automate species identification of three of the world's most important tuna species. The applications of this in fishery monitoring programs globally are almost unlimited, although you don't not even mention this. I suggest you rewrite the paper with this in mind.
Some of the text is not needed and I have done my best with suggesting modifications to text and grammar - the grammar needs to be thoroughly reviewed before any resubmission. Even species names are mis-spelt in the text.
Comments for author File: 
Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
The manuscript of Ou et al. put the attention on a topic of high interest (species identification using artificial intelligence), destined to acquire ever greater importance, especially considering biological management resources and fisheries aspects. The work is clear and well structured, I only suggest a general revision of the English language and to ad a reference. See comments below:
Line 339 quote Bakhshalizadeh et al., 2022: Morphometric Analyses of Phenotypic Plasticity in Habitat Use in Two Caspian Sea Mullet. MDPI, JMSE.
Line 441: Thunnus in italics please.
Line 432: as 441.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Tunas are the basis for important fisheries, and automated identification on ship or in port could provide information useful for fishery management. Ou et al. developed machine learning algorithms that performed well for identifying three species of tuna. Their approach and findings will guide other such studies. That said, the manuscript is hard to follow, especially for readers of this journal that are not strong on machine learning methods. The manuscript will have to be revised to make the methods, analytics, and overall results clearer to the full range of readers. I offer some comments here and also provide a marked manuscript to help with needed revision of the prose. Two matters came up frequently. Rather than “automatic”, the authors should refer to their method as “automated”. Rather than “characteristics”, the respective morphological traits are referred to as “characters”.
Abstract. – At lines 28 and 29, the acronyms KNN and SVM should be defined.
Introduction. – At line 44, since there are many tuna species, the sentence should read that “They have large value…”.
At lines 80 and 81, the respective acronyms must all be defined at first usage.
Methods. – Figure 1 needs a more complete caption. I have written a suggestion on the manuscript document.
At line 149, ROC and AUC must be defined for the reader.
At line 167, we need a supporting citation for the program ‘pyefd’. Similarly, at line 177, we need a supporting citation for program VGG16.
At line 230, not four, but rather three tuna species were the subject of this manuscript.
A sentence at lines 237-238 makes no sense; I’ve marked the manuscript, but I am not confident that I understood the intent of the sentence well enough to offer a useful fix.
Results. – I have no major comments on this section, but have written a number of minor suggested fixes on the manuscript document.
Discussion. – At line 340, many morphological characters are determined not when the fish is hatched (they are not “born”), but rather as they develop, that is, as they undergo the series of morphological transformations over the first few weeks of life.
At line 351, the authors omit mention of a major possible utility of their work. Automated identification of tunas would contribute to precise quantification of the catch, which is highly useful information for fishery management.
References. – I found a few minor errors in the literature citations.
Comments for author File: Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
I thank the authors for the rewrite. Some minor comments to address;
- Line 95 - 'tuna' are not a species. Please provide species name
- line 257 - should 'transform' in this line be 'transformation' in both instances?
I suggest a final proof read before submission
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
