Multimodal skin lesion classification using deep learning
Binary classification
Skin lesion
Contextual image classification
Multiclass classification
DOI:
10.1111/exd.13777
Publication Date:
2018-09-06T01:58:46Z
AUTHORS (3)
ABSTRACT
Abstract While convolutional neural networks (CNNs) have successfully been applied for skin lesion classification, previous studies generally considered only a single clinical/macroscopic image and output binary decision. In this work, we presented method which combines multiple imaging modalities together with patient metadata to improve the performance of automated diagnosis. We evaluated our on classification task comparison as well five class representative real‐world clinical scenario. showed that multimodal classifier outperforms baseline uses macroscopic in both melanoma detection ( AUC 0.866 vs 0.784) multiclass mAP 0.729 0.598). addition, quantitatively diagnosis lesions using dermatoscopic images obtains higher when compared images. performed experiments new data set 2917 cases where each case contains image, metadata.
SUPPLEMENTAL MATERIAL
Coming soon ....
REFERENCES (36)
CITATIONS (230)
EXTERNAL LINKS
PlumX Metrics
RECOMMENDATIONS
FAIR ASSESSMENT
Coming soon ....
JUPYTER LAB
Coming soon ....