Interactive cell segmentation based on phase contrast optics.

Area of Science:

• Biomedical Imaging
• Computational Biology
• Microscopy

Background:

• Cell segmentation is vital for computer-aided analysis of microscopy images.
• Phase contrast microscopy presents challenges due to image artifacts and quality issues.
• Existing methods struggle with accurate cell segmentation in challenging phase contrast images.

Purpose of the Study:

• To develop an interactive cell segmentation scheme for phase contrast microscopy images.
• To address limitations of current methods by leveraging phase retardation features.
• To improve the accuracy and robustness of cell segmentation in complex microscopy data.

Main Methods:

• The proposed method partitions images into phase homogeneous atoms.
• Human annotations are propagated using an affinity graph learned via discrimination analysis.
• An active query strategy selects informative unlabeled atoms for user annotation, enhancing segmentation iteratively.

Main Results:

• The interactive scheme achieves quality cell segmentation results after several rounds of user interaction.
• The approach effectively segments cells exhibiting diverse optical properties.
• Experimental validation demonstrates the robustness of the proposed segmentation method.

Conclusions:

• The interactive cell segmentation scheme based on phase retardation features offers a robust solution for phase contrast microscopy.
• The method successfully integrates user intention with image characteristics for accurate segmentation.
• This approach enhances the foundation for subsequent computer-aided cell image analysis.