A Hybrid Multi-Object Segmentation Framework with Model-Based B-Splines for Microbial Single Cell Analysis

Abstract

In this paper, we propose a hybrid approach for multi-object microbial cell segmentation. The approach combines an MLbased detection with a geometry-aware variational-based segmentation using B-splines that are parametrized based on a geometric model of the cell shape. The detection is done first using YOLOv5. In a second step, each detected cell is segmented individually. Thus, the segmentation only needs to be done on a per-cell basis, which makes it amenable to a variational approach that incorporates prior knowledge on the geometry. Here, the contour of the segmentation is modelled as closed uniform cubic B-spline, whose control points are parametrized using the known cell geometry. Compared to purely ML-based segmentation approaches, which need accurate segmentation maps as training data that are very laborious to produce, our method just needs bounding boxes as training data. Still, the proposed method performs on par with ML-based segmentation approaches usually used in this context. We study the performance of the proposed method on time-lapse microscopy data of Corynebacterium glutamicum.

AI-Powered Microfluidics: Shaping the Future of Phenotypic Drug Discovery

“…leveraged CNN for cell image processing, to enhance cell migration analysis in angiogenic vessels formed in 3D microfluidic environments, achieving an impressive 86.4% accuracy in cell association . Ruzaeva et al introduced an innovative approach that combines machine learning-based detection with geometric-aware segmentation, particularly useful in time-lapse microscopy data analysis of Corynebacterium glutamicum . A study by Liao et al developed a mobile blood acquisition and analysis system that used microfluidic lensless-sensing and CNNs to achieve 98.44% accuracy in cell segmentation and classification.…”

“…102 Ruzaeva et al introduced an innovative approach that combines machine learning-based detection with geometric-aware segmentation, particularly useful in time-lapse microscopy data analysis of Corynebacterium glutamicum. 103 A study by Liao et al developed a mobile blood acquisition and analysis system that used microfluidic lensless-sensing and CNNs to achieve 98.44% accuracy in cell segmentation and classification. This demonstrated how AI can improve cell segmentation in microfluidics, making it possible to conduct precise analysis in situations where resources are limited (Figure 3b).…”

Automated Characterization of Catalytically Active Inclusion Body Production in Biotechnological Screening Systems

“…To separate the main (fully present) cell from parts of cells in the neighborhood included in the corresponding bounding box, a variational B-splines-based segmentation [8] is performed, where the cell shape is modeled as a straight rod (by fixing the curvature parameters d and e in the notation of [8], to zero), which consists of six control points. As the shape of the cell is more flexible than the shape model used in the variational segmentation step, to obtain the fine cell contour, global manually set thresholding (the same threshold value was applied to the entire dataset) is applied to the pixels which belong to the interior of the spline, which results in a binary pixel mask of the cell body.…”

“…The resulting quality metrics for the empirically tuned parameter set (see below), which provides the best scores, are Dice cell = 0.88, Dice CatIB = 0.78. The parameter set includes the cell body and the CatIB thresholds; sizes of the masks for the dilation and erosion, and the weights for the variational spline-based segmentation step [8].…”