Classification of benign and malignant breast tumors based on hybrid level set segmentation

作者：

Highlights：

• Improved GA and CNN provided level set with more efficient initial boundaries compared to SFC.

• Level set parameters were tuned using a dynamic training procedure adaptively and automatically.

• MLP produces the highest classification accuracy among other classifiers.

• Adaptive segmentation methods achieved higher performance than that of the first proposed methods.

摘要

•Improved GA and CNN provided level set with more efficient initial boundaries compared to SFC.•Level set parameters were tuned using a dynamic training procedure adaptively and automatically.•MLP produces the highest classification accuracy among other classifiers.•Adaptive segmentation methods achieved higher performance than that of the first proposed methods.

论文关键词：Region growing,Cellular neural network,Level set method,Genetic algorithm,Memetic algorithm,Artificial neural network

论文评审过程：Received 14 May 2015, Revised 10 October 2015, Accepted 11 October 2015, Available online 26 October 2015, Version of Record 18 November 2015.

论文官网地址：https://doi.org/10.1016/j.eswa.2015.10.011