Related Experiment Video
Updated: Jun 26, 2026
09:31
Automated Quantification of Hematopoietic Cell – Stromal Cell Interactions in Histological Images of Undecalcified Bone
Published on: April 8, 2015
[A method used in myeloma cell cloning spot image segmentation]
Yu-ya Yao1, Zhuang-zhi Yan, Yu Chen
1School of Communication and Information Engineering, Shanghai University Institute of Biomedical Engineering.
Summary
This study introduces an interactive image segmentation method for multiple myeloma cloning spots. The graph cut-based approach effectively separates cellular structures, improving diagnostic accuracy.
Area of Science:
- Computational Biology
- Medical Imaging Analysis
- Image Processing
Context:
- Accurate segmentation of multiple myeloma cloning spots is crucial for disease assessment.
- Existing methods may lack the interactivity and precision required for complex cellular structures.
- Automated image analysis tools are increasingly vital in hematological research.
Purpose:
- To present a novel interactive image segmentation technique for multiple myeloma cloning spots.
- To leverage graph cut theory and K-means clustering for efficient image partitioning.
- To enable user-guided refinement for improved segmentation accuracy.
Summary:
- The method utilizes graph cuts, initializing with K-means clustering to define object and background seeds.
- An energy minimization process segments the image, followed by morphological operations (erosion and dilation).
- An interactive refinement tool allows users to correct misclassified pixels, enhancing segmentation precision.
Impact:
- Provides a more accurate and efficient tool for analyzing multiple myeloma cell cultures.
- Facilitates improved quantitative analysis in hematological research and diagnostics.
- Demonstrates satisfactory performance based on subjective and RUMA evaluation criteria.
