Related Experiment Video
Updated: Sep 4, 2025
Area-based Image Analysis Algorithm for Quantification of Macrophage-fibroblast Cocultures
Published on: February 15, 2022
Marker-controlled watershed with deep edge emphasis and optimized H-minima transform for automatic segmentation of
Tuomas Kaseva1, Bahareh Omidali2, Eero Hippeläinen2,3
1HUS Medical Imaging Center, Radiology, Helsinki University Hospital and University of Helsinki, P.O. Box 340, FI-00290, Helsinki, Finland.
Background:
The segmentation of 3D cell nuclei is essential in many tasks, such as targeted molecular radiotherapies (MRT) for metastatic tumours, toxicity screening, and the observation of proliferating cells. In recent years, one popular method for automatic segmentation of nuclei has been deep learning enhanced marker-controlled watershed transform. In this method, convolutional neural networks (CNNs) have been used to create nuclei masks and markers, and the watershed algorithm for the instance segmentation. We studied whether this method could be improved for the segmentation of densely cultivated 3D nuclei via developing multiple system configurations in which we studied the effect of edge emphasizing CNNs, and optimized H-minima transform for mask and marker generation, respectively.
Results:
The dataset used for training and evaluation consisted of twelve in vitro cultivated densely packed 3D human carcinoma cell spheroids imaged using a confocal microscope. With this dataset, the evaluation was performed using a cross-validation scheme. In addition, four independent datasets were used for evaluation. The datasets were resampled near isotropic for our experiments. The baseline deep learning enhanced marker-controlled watershed obtained an average of 0.69 Panoptic Quality (PQ) and 0.66 Aggregated Jaccard Index (AJI) over the twelve spheroids. Using a system configuration, which was otherwise the same but used 3D-based edge emphasizing CNNs and optimized H-minima transform, the scores increased to 0.76 and 0.77, respectively. When using the independent datasets for evaluation, the best performing system configuration was shown to outperform or equal the baseline and a set of well-known cell segmentation approaches.
Conclusions:
The use of edge emphasizing U-Nets and optimized H-minima transform can improve the marker-controlled watershed transform for segmentation of densely cultivated 3D cell nuclei. A novel dataset of twelve spheroids was introduced to the public.
Insights
Improving 3D cell nuclei segmentation using edge emphasizing convolutional neural networks (CNNs) and optimized H-minima transform enhances marker-controlled watershed segmentation for dense cell cultures.
Area of Science:
- * Biomedical imaging
- * Computational biology
- * Machine learning for image analysis
Background:
- * Accurate 3D cell nuclei segmentation is crucial for applications like cancer radiotherapy, toxicity screening, and cell proliferation studies.
- * Deep learning-enhanced marker-controlled watershed transform is a common method for automatic nuclei segmentation.
- * This study investigates improvements for segmenting densely packed 3D nuclei.
Purpose of the Study:
- * To enhance the marker-controlled watershed transform for segmenting densely cultivated 3D cell nuclei.
- * To evaluate the impact of edge-emphasizing CNNs and optimized H-minima transform on segmentation performance.
Main Methods:
- * Development and testing of multiple system configurations for 3D cell nuclei segmentation.
- * Utilization of edge-emphasizing convolutional neural networks (CNNs) for mask and marker generation.
- * Application of optimized H-minima transform for mask and marker refinement.
- * Evaluation using a dataset of 12 in vitro cultivated 3D human carcinoma cell spheroids and four independent datasets.
Main Results:
- * The baseline deep learning enhanced marker-controlled watershed achieved an average Panoptic Quality (PQ) of 0.69 and Aggregated Jaccard Index (AJI) of 0.66.
- * The improved system configuration using 3D-based edge emphasizing CNNs and optimized H-minima transform increased PQ to 0.76 and AJI to 0.77.
- * The optimized configuration outperformed or matched established cell segmentation methods on independent datasets.
Conclusions:
- * Edge-emphasizing U-Nets combined with optimized H-minima transform significantly improve marker-controlled watershed segmentation for dense 3D cell nuclei.
- * A novel dataset comprising twelve spheroids has been made publicly available.
More Related Videos
14:28Substructure Analyzer: A User-Friendly Workflow for Rapid Exploration and Accurate Analysis of Cellular Bodies in Fluorescence Microscopy Images
Published on: July 15, 2020
12:08From Voxels to Knowledge: A Practical Guide to the Segmentation of Complex Electron Microscopy 3D-Data
Published on: August 13, 2014