List of Papers By topics Author List
| Paper Info | Reviews | Meta-review | Author Feedback | Post-Rebuttal Meta-reviews |
Authors
Yan-Jie Zhou, Wei Liu, Yuan Gao, Jing Xu, Le Lu, Yuping Duan, Hao Cheng, Na Jin, Xiaoyong Man, Shuang Zhao, Yu Wang
Abstract
Skin diseases are among the most prevalent health issues, and accurate computer-aided diagnosis methods are of importance for both dermatologists and patients. However, most of the existing methods overlook the essential domain knowledge required for skin disease diagnosis. A novel multi-task model, namely DermImitFormer, is proposed to fill this gap by imitating dermatologists’ diagnostic procedures and strategies. Through multi-task learning, the model simultaneously predicts body parts and lesion attributes in addition to the disease itself, enhancing diagnosis accuracy and improving diagnosis interpretability. The designed lesion selection module mimics dermatologists’ zoom-in action, effectively highlighting the local lesion features from noisy backgrounds. Additionally, the presented cross-interaction module explicitly models the complicated diagnostic reasoning between body parts, lesion attributes, and diseases. To provide a more robust evaluation of the proposed method, a large-scale clinical image dataset of skin diseases with significantly more cases than existing datasets has been established. Extensive experiments on three different datasets consistently demonstrate the state-of-the-art recognition performance of the proposed approach.
Link to paper
DOI: https://doi.org/10.1007/978-3-031-43987-2_20
SharedIt: https://rdcu.be/dnwJC
Link to the code repository
N/A
Link to the dataset(s)
N/A
Reviews
Review #3
- Please describe the contribution of the paper
The paper proposes a multi-task model, called DermImitFormer, for the differential diagnosis of skin diseases. The model comprises three modules, including multi-task learning, a lesion selection module that highlights local lesion features, and a cross-interaction module. The performance of the proposed model is evaluated on three datasets.
- Please list the main strengths of the paper; you should write about a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.
- The DermImitFormer model is proposed, which includes multi-task learning strategy, a lesion selection module for learning distinct features and a cross-interaction module for fusing three different representations.
- The authors created a new dataset containing 57,246 images, representing 49 skin diseases across 15 body parts and 27 lesion attributes.
- Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.
The ablation studies in the paper suggest that there is only slight improvement in performance when using the cross-interaction module (CIM) instead of the Concat fusion method.
- Please rate the clarity and organization of this paper
Good
- Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance
Although the paper provides some details that can aid reproducibility, it would be beneficial to release the code and dataset to make the research more reproducible.
- Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://conferences.miccai.org/2023/en/REVIEWER-GUIDELINES.html
The paper makes a significant contribution to the field of skin disease diagnosis and has great potential for practical applications. However, to enhance reproducibility, it would be beneficial for the authors to release the code and dataset. It might also be helpful to provide information about the model’s computational time. Additionally, Figure 3 could be improved to facilitate better understanding at a glance.
- Rate the paper on a scale of 1-8, 8 being the strongest (8-5: accept; 4-1: reject). Spreading the score helps create a distribution for decision-making
6
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
The paper proposes a multi-task model for skin disease diagnosis, which includes multi-task learning, a lesion selection module, and a cross-interaction module. The authors also establish a new dataset and perform a reasonable evaluation on three different datasets.
- Reviewer confidence
Confident but not absolutely certain
- [Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed
N/A
- [Post rebuttal] Please justify your decision
N/A
Review #1
- Please describe the contribution of the paper
In this paper, a novel multi-task model DermImitFormer is proposed to imitate dermatologists’ diagnostic processes. And the authors established a large-scale clinical image dataset of skin diseases.
- Please list the main strengths of the paper; you should write about a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.
The innovation of this article is to use dermatologists’ domain knowledge by mimicking their subjective diagnostic procedures. At the same time, a large-scale clinical image dataset of skin diseases with significantly more cases than existing datasets has been established.
- Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.
In the paper, the limitations of the comparative method can be properly explained, which can better highlight the superiority of the method.
- Please rate the clarity and organization of this paper
Very Good
- Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance
The paper has good reproducibility for it has a clear model discussion, sufficient formula derivation and implementation details.
- Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://conferences.miccai.org/2023/en/REVIEWER-GUIDELINES.html
In the paper, the limitations of the comparative method can be properly explained, which can better highlight the superiority of the method.
- Rate the paper on a scale of 1-8, 8 being the strongest (8-5: accept; 4-1: reject). Spreading the score helps create a distribution for decision-making
6
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
The article is well structured and the methodology of this article is clearly explained. Moerover, experimental results demostrate the advanced performance of the method in this paper, this score is given.
- Reviewer confidence
Confident but not absolutely certain
- [Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed
N/A
- [Post rebuttal] Please justify your decision
N/A
Review #2
- Please describe the contribution of the paper
In this article, the author generates a novel multi-task model called DermImitFormer, which can imitate dermatologists’ diagnostic procedures and strategies for accurate differential diagnosis of skin diseases in clinical images. What’s more, the author also creates a novel clinical dataset of skin diseases called Derm-49.
- Please list the main strengths of the paper; you should write about a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.
The model, which mainly uses vision transformer for the main structure and adds unique information about skin diseases such as body part and attribute to amend the parameters of the transformer layer, gets the highest Acc and F1 in comparison.
- Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.
1.The proposed method is suitable for the dataset created in the article, but most of the datasets do not have the corresponding labels such as attributes, which can limit the multitasking feature of the model, thus affecting the generalizability of the model. 2.In the part of optimization, eq. 6 seems like a binary cross entropy loss, yet La and Lb is a multi-class loss. However, the article has no adequate explanation for this.
- Please rate the clarity and organization of this paper
Good
- Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance
This paper can be reproduced.
- Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://conferences.miccai.org/2023/en/REVIEWER-GUIDELINES.html
There are few grammatical errors in this article and the structure of this article is suitable for MICCAI.
- Rate the paper on a scale of 1-8, 8 being the strongest (8-5: accept; 4-1: reject). Spreading the score helps create a distribution for decision-making
6
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
The whole article is smoothly written and the experiments are well designed. The model simulates the behavior of doctors during diagnosis, which is quite persuasive. The proposed LSM model can select dermatosis area, which is helpful for diagnosis.
- Reviewer confidence
Very confident
- [Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed
N/A
- [Post rebuttal] Please justify your decision
N/A
Primary Meta-Review
- Please provide your assessment of this work, taking into account all reviews. Summarize the key strengths and weaknesses of the paper and justify your recommendation. In case you deviate from the reviewers’ recommendations, explain in detail the reasons why. In case of an invitation for rebuttal, clarify which points are important to address in the rebuttal.
In this paper, by simulating the diagnosis process of clinicians, a multi-task network model is proposed for the intelligent diagnosis of skin diseases. At the same time, a large-scale dermatology clinical image dataset with significantly more cases than existing datasets is established. The starting point and method setting of the article are innovative. Combined with the opinions of multiple reviewers, to enhance the algorithm’s reproducibility, the author’s release of codes and data sets will be beneficial to the development of intelligent skin disease diagnosis.
Author Feedback
N/A