Related Experiment Video
Updated: Sep 27, 2025
10:55
Live Imaging Followed by Single Cell Tracking to Monitor Cell Biology and the Lineage Progression of Multiple Neural Populations
Published on: December 16, 2017
8.8K
DMNet: Dual-Stream Marker Guided Deep Network for Dense Cell Segmentation and Lineage Tracking
Rina Bao1, Noor M Al-Shakarji1, Filiz Bunyak1
1University of Missouri-Columbia, MO 65211, USA.
... IEEE International Conference on Computer Vision Workshops. IEEE International Conference on Computer Vision
|April 7, 2022
Summary
This study introduces a dual-stream marker-guided network (DMNet) for precise cell segmentation and tracking in microscopy images. The method excels at handling challenging cases like dense, touching, and deforming cells, improving biomedical research and diagnostics.
Area of Science:
- Biomedical imaging
- Computational biology
- Machine learning for cell analysis
Background:
- Accurate cell segmentation and tracking in microscopy are crucial for clinical diagnostics and research.
- Challenges include segmenting dense, touching, and deforming cells with indistinct boundaries in low signal-to-noise images.
Purpose of the Study:
- To develop a robust algorithm for accurate segmentation and tracking of cells in microscopy videos.
- To address the limitations of current methods in handling complex cellular structures and image qualities.
Main Methods:
- A dual-stream marker-guided network (DMNet) was developed for cell segmentation, featuring separate marker-detection and mask-prediction streams.
- A distance map penalty function was employed to focus training on challenging touching and nearby cells.
- The M2Track tracking-by-detection approach with multi-step data association was utilized for multi-object cell tracking, incorporating track-to-cell and track-to-track association.
Main Results:
- The combined DMNet and M2Track algorithm demonstrated high performance in segmenting and tracking diverse cell types.
- The approach successfully handled dense, touching, and deforming cells in challenging microscopy images.
- Achieved multiple top three rankings in the IEEE ISBI 2021 6th Cell Tracking Challenge (CTC-6).
Conclusions:
- The proposed dual-stream network and tracking method offer a significant advancement for automated cell analysis in microscopy.
- This technique has proven effective in complex scenarios and competitive benchmarks, showing potential for clinical and research applications.
