Related Experiment Videos
Algorithms for cytoplasm segmentation of fluorescence labelled cells
Carolina Wählby1, Joakim Lindblad, Mikael Vondrus
1Centre for Image Analysis at Uppsala University, Uppsala, Sweden. carolina@cb.uu.se
Summary
This study introduces a novel automated algorithm for cell cytoplasm segmentation in fluorescence microscopy. The method achieves high accuracy, between 89-97%, for segmenting Chinese Hamster Ovary (CHO) cells.
Area of Science:
- Cell Biology
- Biomedical Imaging
- Computational Biology
Background:
- Accurate cell segmentation is crucial for quantitative analysis in cytometry.
- While nuclear segmentation is straightforward, cytoplasm segmentation remains challenging due to complex cell shapes and clustering.
Purpose of the Study:
- To develop and validate a fully automatic image analysis algorithm for accurate cell cytoplasm segmentation.
- To improve the efficiency and reliability of cell segmentation in fluorescence microscopy.
Main Methods:
- A novel algorithm combining image pre-processing, general segmentation, merging, and quality measurement.
- Utilized statistical analysis of shape features for a feedback-driven splitting step to separate clustered cells.
- Trained the algorithm on representative images for fully automatic subsequent segmentation.
Main Results:
- The algorithm demonstrated robust performance in segmenting cytoplasm of Chinese Hamster Ovary (CHO) cells.
- Achieved a high accuracy rate of 89% to 97% compared to manual segmentation.
- The developed quality measurement and splitting steps effectively handled challenging clustered cell scenarios.
Conclusions:
- The proposed automated method significantly enhances the accuracy and efficiency of cell cytoplasm segmentation.
- This algorithm offers a reliable tool for quantitative cell analysis in fluorescence microscopy applications.
- The approach is adaptable for segmenting other cell types under similar imaging conditions.