Hierarchical partial matching and segmentation of interacting cells.

Tracking and segmenting interacting cells in phase-contrast microscopy remains a challenge due to cell deformation and overlapping boundaries. Prior research has shown that standard segmentation methods struggle when cells touch or merge. No prior work had resolved the issue of partial boundary matching after collision events. This gap motivated the development of a new approach that handles complex interactions. Existing techniques often fail to maintain accurate segmentation across multiple frames. That uncertainty drove the need for a method that can track individual cells even during complex interactions. The problem is particularly relevant in developmental biology and tissue engineering. This paper's contribution lies in addressing the limitations of current methods in such scenarios.

The aim of this work is to develop an automated method for tracking and segmenting living cells in phase-contrast image sequences. The specific problem is the difficulty in maintaining accurate segmentation when cells deform or interact. The motivation stems from the limitations of existing techniques in handling overlapping boundaries. The researchers propose a new approach that can handle complex cell interactions. The study focuses on scenarios where a single boundary encloses multiple touching cells. The goal is to cut and reconnect these boundaries accurately. The method is designed to work across thousands of images with multiple interactions. This approach is intended to improve accuracy in cell tracking and segmentation.

The researchers formulate the problem as a many-to-one elastic partial matching task between closed curves. They introduce Double Cyclic Dynamic Time Warping to handle collision events. This method allows for matching boundaries that have been merged into a single curve. The approach converts the partial-curve matching problem into a shortest path problem. They solve this efficiently using a shortest path tree. The same algorithm is used to fill gaps between segments of target curves. The method is tested on phase-contrast image sequences with interacting cells. The results are validated across thousands of images to ensure accuracy.

The method successfully tracks and segments interacting cells in phase-contrast image sequences. It handles complex interactions such as cell deformation and overlapping boundaries. The researchers demonstrate the effectiveness of their approach across 8068 images. The method maintains accurate recognition of individual cell boundaries. It uses a shortest path algorithm to reconnect segments after collision events. The Double Cyclic Dynamic Time Warping improves boundary matching accuracy. The results show improved performance compared to prior methods. The approach is efficient and scalable for large datasets.

The authors propose a method that improves tracking and segmentation of interacting cells in phase-contrast microscopy. Their approach addresses the challenge of maintaining accurate boundaries after collision events. The use of Double Cyclic Dynamic Time Warping is a novel contribution. The method converts partial-curve matching into a shortest path problem. This allows for efficient and accurate segmentation of overlapping cells. The results demonstrate improved performance across a large dataset. The approach is effective in handling complex interactions. The authors suggest that their method can be applied to various biological imaging scenarios.