ABSTRACT
Introduction: Acne detection is critical in dermatology, focusing on quality control of acne imagery, precise segmentation, and grading. Traditional research has been limited, typically concentrating on singular aspects of acne detection. Methods: We propose a multi-task acne detection method, employing a CenterNet-based training paradigm to develop an advanced detection system. This system collects acne images via smartphones and features multi-task capabilities for detecting image quality and identifying various acne types. It differentiates between noninflammatory acne, papules, pustules, nodules, and provides detailed delineation for cysts and post-acne scars. Results: The implementation of this multi-task learning-based framework in clinical diagnostics demonstrated an 83% accuracy in lesion categorization, surpassing ResNet18 models by 12%. Furthermore, it achieved a 76% precision in lesion stratification, outperforming dermatologists by 16%. Discussion: Our framework represents a advancement in acne detection, offering a comprehensive tool for classification, localization, counting, and precise segmentation. It not only enhances the accuracy of remote acne lesion identification by doctors but also clarifies grading logic and criteria, facilitating easier grading judgments.
ABSTRACT
Skin diseases pose significant challenges in China. Internet health forums offer a platform for millions of users to discuss skin diseases and share images for early intervention, leaving large amount of valuable dermatology images. However, data quality and annotation challenges limit the potential of these resources for developing diagnostic models. In this study, we proposed a deep-learning model that utilized unannotated dermatology images from diverse online sources. We adopted a contrastive learning approach to learn general representations from unlabeled images and fine-tuned the model on coarsely annotated images from Internet forums. Our model classified 22 common skin diseases. To improve annotation quality, we used a clustering method with a small set of standardized validation images. We tested the model on images collected by 33 experienced dermatologists from 15 tertiary hospitals and achieved a 45.05% top-1 accuracy, outperforming the published baseline model by 3%. Accuracy increased with additional validation images, reaching 49.64% with 50 images per category. Our model also demonstrated transferability to new tasks, such as detecting monkeypox, with a 61.76% top-1 accuracy using only 50 additional images in the training process. We also tested our model on benchmark datasets to show the generalization ability. Our findings highlight the potential of unannotated images from online forums for future dermatology applications and demonstrate the effectiveness of our model for early diagnosis and potential outbreak mitigation.
ABSTRACT
Background: Due to the lower reliability of laboratory tests, skin diseases are more suitable for diagnosis with AI models. There are limited AI dermatology diagnostic models combining images and text; few of these are for Asian populations, and few cover the most common types of diseases. Methods: Leveraging a dataset sourced from Asia comprising over 200,000 images and 220,000 medical records, we explored a deep learning-based system for Dual-channel images and extracted text for the diagnosis of skin diseases model DIET-AI to diagnose 31 skin diseases, which covers the majority of common skin diseases. From 1 September to 1 December 2021, we prospectively collected images from 6,043 cases and medical records from 15 hospitals in seven provinces in China. Then the performance of DIET-AI was compared with that of six doctors of different seniorities in the clinical dataset. Results: The average performance of DIET-AI in 31 diseases was not less than that of all the doctors of different seniorities. By comparing the area under the curve, sensitivity, and specificity, we demonstrate that the DIET-AI model is effective in clinical scenarios. In addition, medical records affect the performance of DIET-AI and physicians to varying degrees. Conclusion: This is the largest dermatological dataset for the Chinese demographic. For the first time, we built a Dual-channel image classification model on a non-cancer dermatitis dataset with both images and medical records and achieved comparable diagnostic performance to senior doctors about common skin diseases. It provides references for exploring the feasibility and performance evaluation of DIET-AI in clinical use afterward.
ABSTRACT
The primary focus of this study is to evaluate the impact of various levels of education on CO2 emissions in China. Moreover, the study also tested the EKC hypothesis for different levels of education and economic development. The analysis employed disaggregate and aggregate data for education that included enrollment at primary, secondary, and tertiary levels and the average year of schooling. For empirical analysis, we employed an error correction model and bounds testing approach to cointegration. The results of the study provided some useful information both in the short and long run. All the proxies of education positively impact CO2 emissions at the initial level both in the short and long run; however, when we take the square of these variables, the effects of education on CO2 emissions become negative. Similarly, the impact of economic growth on CO2 emissions is positive in the short and long run, and the square of economic growth on CO2 emissions is negative, supporting the EKC hypothesis. China should increase investment in human capital that promotes green growth and environmental quality.
