All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
The paper can be accepted.
[# PeerJ Staff Note - this decision was reviewed and approved by Shawn Gomez, a PeerJ Section Editor covering this Section #]
-
-
-
Thanks for the revision. The changes to the paper and the final version are generally adequate. I have no further revision requirements. Best regards.
**PeerJ Staff Note:** Please ensure that all review, editorial, and staff comments are addressed in a response letter and that any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate.
**Language Note:** When you prepare your next revision, please either (i) have a colleague who is proficient in English and familiar with the subject matter review your manuscript, or (ii) contact a professional editing service to review your manuscript. PeerJ can provide language editing services - you can contact us at [email protected] for pricing (be sure to provide your manuscript number and title). – PeerJ Staff
The sentencing has improved since the first version. The revised version is clearer and easier to understand.
Yes, the Methods are sufficiently described, and the corresponding images are well documented.
Yes, experiments are explained well and evaluation is satisfactory.
The revised version explains the rationale of using the Domain Adversarial Neural Network and Transfer learning in detecting the emotions of the realistic and figurative art images. The experiments are satisfactory, and the authors have generated relevant images to explain the approach and proposed solution.
-
-
-
Thanks for the revision. Please ensure that the missing evaluation metrics referenced in step eight of my initial comments are obtained and reported, as their inclusion is essential for a complete and rigorous assessment of the proposed method.
8. Certain metrics have been obtained in terms of evaluation metrics. However, for the classification results to be analyzed more accurately and the proposed model to be more prominent, the missing metrics must definitely be obtained. For a comprehensive and balanced evaluation of the model’s performance, it is essential to compute and report all relevant metrics without omission, including classification accuracy, precision, recall, F1-score, macro/micro averages, Hamming loss, subset accuracy, confusion matrix, ROC-AUC, mean average precision (mAP), Cohen’s Kappa, and Matthews Correlation Coefficient (MCC); therefore, any missing metrics from this list should be included in the study.
**PeerJ Staff Note:** Please ensure that all review and editorial comments are addressed in a response letter and that any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate.
**Language Note:** The review process has identified that the English language must be improved. PeerJ can provide language editing services - please contact us at [email protected] for pricing (be sure to provide your manuscript number and title). Alternatively, you should make your own arrangements to improve the language quality and provide details in your response letter. – PeerJ Staff
Improve grammar in general. Minor issues such as: Firstly, we introduce the attention
mechanism (AM) to further highlight the key emotional cues of the subject area by calculating the
importance weights of different regions for the sentiment classification task, which can capture
the emotional image features more accurately.
can be written as:
Firstly, we introduce the attention mechanism (AM) to further highlight the key emotional cues of the subject area by calculating the importance weights of different regions for the sentiment classification task, which can more accurately capture the emotional image features.
Methods
The method used in the research paper is well written, outlining a standard pipeline that leverages the convolutional neural network with an attention mechanism and transfer learning to extract salient features and analyze for expressions, in this case, emotions.
The domain adaptation approach described in section 3.3 is a Domain-Adversarial Neural Network (DANN). This should be called out and cite the original work. Also, make sure to cite wherever this is applicable, for instance under Section 5, conclusion.
In the The framework of AMSF diagram (Fig1), call out the type of attention mechanism for clarity
I recommend listing he method names in Fig 5 for easier understanding.
It is important to describe the type of artistic images used for clarity in the following text under Section 5: the conclusion & limitations:
"The experiments conclude that our method achieves a 92.4% mAP score and 98.9% accuracy in the artistic image sentiment classification, which fully proves its effectiveness and reliability in practical applications. "
The final section contains a comprehensive presentation of all comments.
The final section contains a comprehensive presentation of all comments.
The final section contains a comprehensive presentation of all comments.
Peer Review Report for PeerJ Computer Science
Manuscript Title: Art image emotional classification by domain adaptation and transfer learning
1. This study presents an effective approach for interpreting the emotional content of artworks by using a domain-adaptive method that combines attention-based feature extraction and transfer learning for accurate emotion classification in artistic images.
2. It is recommended that the abstract section of the study be further elaborated to provide a clearer overview of the research problem, methodology, key findings, and the significance of the results.
3. In the introduction, the importance of the subject, the difficulties encountered for emotional classification of art images, and the contributions of the study are mentioned at a basic level. At the end of this section, the main contributions of the study to the literature, its originality, and its difference from the literature should be stated more clearly in a more detailed and itemized manner.
4. Although the artificial intelligence studies in the literature related to image sentiment classification are stated, they are very limited. A more detailed analysis should definitely be made in this section. In addition, an in-depth literature table consisting of columns such as dataset, originality, deficiencies, and literature contributions should be added. After this, the advantages of this study compared to the literature and the deficiencies it will eliminate should be clearly detailed.
5. When the dataset, amount, type, and data augmentation and preprocessing steps used in the study are carefully examined within the scope of the study, it is observed that they are sufficient. However, sample dataset images for each class and sample images for each class after the data augmentation/preprocessing steps should be added.
6. Using domain adaptation for image sentiment classification processes has increased the quality of the study. However, it is recommended that the study be classified with a state-of-the-art model and the results be compared so that it can be more prominent in the literature.
7. It is positive that the parameters related to the model hyperparameters are given together with the table. However, it should be explained more clearly how these metrics were determined and whether different experiments were made. In addition, it is recommended that all missing parameter types and values be added to the table with their reasons.
8. Certain metrics have been obtained in terms of evaluation metrics. However, in order for the classification results to be analyzed more accurately and the proposed model to be more prominent, the missing metrics must definitely be obtained. For a comprehensive and balanced evaluation of the model’s performance, it is essential to compute and report all relevant metrics without omission, including classification accuracy, precision, recall, F1-score, macro/micro averages, Hamming loss, subset accuracy, confusion matrix, ROC-AUC, mean average precision (mAP), Cohen’s Kappa, and Matthews Correlation Coefficient (MCC); therefore, any missing metrics from this list should be included in the study.
In conclusion, the study is interesting, but all the above sections should be addressed completely and in great detail in terms of literature analysis, originality, and analysis of the results.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.