Convolutional neural network approach for the automated identification of in cellulo crystals

This repository contains all files required to complement our manuscript entitled "Convolutional neural network approach for the automated identification of in cellulo crystals" published in the Journal of Applied Crystallography (A. Kardoost, R. Schönherr, C. Deiter, L. Redecke, K. Lorenzen, J. Schulz and I. de Diego (2024). J.Appl. Cryst. 57, https://doi.org/10.1107/S1600576724000682). In this work we make use of Mask R-CNN, a Convolutional Neural Network (CNN) based instance segmentation method, for the identification of crystals growing in living insect cells, using conventional bright field images.

本仓库包含补充我们发表于《应用结晶学杂志（Journal of Applied Crystallography）》（2024年，A. Kardoost、R. Schönherr、C. Deiter、L. Redecke、K. Lorenzen、J. Schulz与I. de Diego，DOI: 10.1107/S1600576724000682）的题为《基于卷积神经网络的细胞内晶体自动化识别方法》的手稿所需的全部文件。本研究采用基于卷积神经网络(Convolutional Neural Network, CNN)的实例分割方法Mask R-CNN，通过常规明场图像实现对活体昆虫细胞内生长晶体的识别。