Efficient Column Generation for Cell Detection and Segmentation
This provides a fast and accurate solution for biologists analyzing cell microscopy images, though it appears incremental as it builds on existing column generation techniques.
The paper tackles instance segmentation of crowded cells in biological images by formulating it as an integer program and solving it with a column generation method that uses exact optimization and odd set inequalities. The algorithm achieves or exceeds state-of-the-art accuracy on three microscopy datasets with hundreds of cells each.
We study the problem of instance segmentation in biological images with crowded and compact cells. We formulate this task as an integer program where variables correspond to cells and constraints enforce that cells do not overlap. To solve this integer program, we propose a column generation formulation where the pricing program is solved via exact optimization of very small scale integer programs. Column generation is tightened using odd set inequalities which fit elegantly into pricing problem optimization. Our column generation approach achieves fast stable anytime inference for our instance segmentation problems. We demonstrate on three distinct light microscopy datasets, with several hundred cells each, that our proposed algorithm rapidly achieves or exceeds state of the art accuracy.