Multi-feature representation for burn depth classification via burn images
作者：
Highlights：
• We proposed a multi-feature representation method for detecting burn depth.
• Three different features: color, texture and latent are extracted from burn images.
• The latent feature is extracted by a stacked sparse autoencoder for the first time.
• The proposed method shows effectiveness on a burn image dataset.
• The proposed approach obtains the best results compared to other popular methods.
摘要
•We proposed a multi-feature representation method for detecting burn depth.•Three different features: color, texture and latent are extracted from burn images.•The latent feature is extracted by a stacked sparse autoencoder for the first time.•The proposed method shows effectiveness on a burn image dataset.•The proposed approach obtains the best results compared to other popular methods.
论文关键词：Burn,Burn depth,Classification,Image processing,Multiple features
论文评审过程：Received 8 June 2020, Revised 13 May 2021, Accepted 22 June 2021, Available online 27 June 2021, Version of Record 8 July 2021.
论文官网地址：https://doi.org/10.1016/j.artmed.2021.102128
