To enable earlier detection of MPXV infection, we developed an image-based deep convolutional neural network (named MPXV-CNN) for the identification of the characteristic skin lesions caused by MPXV. Undetected infection and delayed isolation of infected individuals are key factors driving the monkeypox virus (now termed mpox virus or MPXV) outbreak. We used dupeGuru 4.31 to identify duplicate images which is available at. The following packages were used which can be installed with the python package installer (pip): pytorch 1.12.0, fastai 2.7.7,scikit-image 0.19.3, python 3.7.13, torchvision 0.13.0, cudatoolkit 11.6.0, matplotlib 3.5.2. The code for training the MPXV-CNN is available at. The SHAP library used for explainability in this study is available at. The pretrained ResNet34 architecture used for the MPXV-CNN in this work is publicly available within the FastAI framework. The deep-learning framework (FastAI v2) used in this study is available at.
MPXV images of the prospective cohort from the Stanford University Medical Center and the Esteva dataset are nonpublic and cannot be shared. Social media references are available upon request. PAD-UFES-20: /datasets/zr7vgbcyr2/1įitzpatrick 17k: /mattgroh/fitzpatrick17k The images and metadata of datasets can be obtained from the following addresses: This study used publicly available data from publications of the scientific literature, dermatological repositories, news articles and social media.Ī bibliography of sources with MPXV skin lesion images was provided as Supplementary Note 1.ĭermatological repositories with non-MPXV images can be accessed using the following addresses:ĭanderm: danderm-pdv.is.kkh.dk DermIS: HDA: DermNet: / DermNet NZ: Ī list of URLs to cleaned non-MPXV skin disease images of Danderm, DermIS, HDA, was provided as Supplementary Note 2.
0 kommentar(er)
