Related Experiment Video
Updated: Jun 2, 2026
08:40
Quantitation of Protein Expression and Co-localization Using Multiplexed Immuno-histochemical Staining and Multispectral Imaging
Published on: April 8, 2016
Partitioning histopathological images: an integrated framework for supervised color-texture segmentation and cell
Hui Kong1, Metin Gurcan, Kamel Belkacem-Boussaid
1Department of Biomedical Informatics, Ohio State University, Columbus, OH 43201, USA. tom.hui.kong@gmail.com
IEEE Transactions on Medical Imaging
|April 14, 2011
Summary
This study introduces a new framework for analyzing histopathological images, improving cell segmentation and splitting for accurate lymphoma grading. The method enhances feature quantification in medical imaging analysis.
Area of Science:
- Digital Pathology
- Computational Biology
- Image Analysis
Background:
- Accurate quantitative analysis of histopathological images is crucial for tasks like lymphoma grading.
- Current methods often rely on single-cell feature quantification, which can be challenging with complex cell structures.
Purpose of the Study:
- To develop an integrated framework for improved cell segmentation and touching-cell splitting in histopathological images.
- To enhance the accuracy of feature quantification for quantitative analysis in medical imaging.
Main Methods:
- A novel supervised cell-image segmentation algorithm using color-texture features extracted via local Fourier transform (LFT) in a discriminant color space (MDC).
- An efficient LFT extraction algorithm utilizing image shifting and integral techniques.
- A new touching-cell splitting method based on radial symmetry and Fourier shape descriptors.
Main Results:
- The proposed segmentation algorithm outperformed existing methods, including superpixel, graph-cut, and mean-shift, on follicular lymphoma images.
- The touching-cell splitting method achieved a low total error rate of 5.25% per image, with evaluations for under-splitting, over-splitting, and encroachment errors.
Conclusions:
- The integrated framework provides a robust and efficient solution for cell segmentation and splitting in complex histopathological images.
- This approach significantly improves the accuracy of quantitative analysis for applications such as cancer grading.
