Neuron cell count with deep learning in highly dense hippocampus images

作者：

Highlights：

• Innovative way to treat highly dense hippocampal images when they are of low quality.

• Better pixel characterization reduces model size and increases performance in model.

• Deep learning models increase their performance with proposed imaging procedure.

• Metrics show that the procedure is reliable for counting cells on histological images.

摘要

•Innovative way to treat highly dense hippocampal images when they are of low quality.•Better pixel characterization reduces model size and increases performance in model.•Deep learning models increase their performance with proposed imaging procedure.•Metrics show that the procedure is reliable for counting cells on histological images.

论文关键词：Histological images,Deep learning,Neuron cell count,Hematoxylin Eosin

论文评审过程：Received 2 November 2021, Revised 4 July 2022, Accepted 4 July 2022, Available online 8 July 2022, Version of Record 17 July 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.118090