Detection of mitosis within a stem cell population of high cell confluence in phase-contrast microscopy images

Abstract

Computer vision analysis of cells in phase-contrast microscopy images enables long-term continuous monitoring of live cells, which has not been feasible using the existing cellular staining methods due to the use of fluorescence reagents or fixatives. In cell culture analysis, accurate detection of mitosis, or cell division, is critical for quantitative study of cell proliferation. In this work, we present an approach that can detect mitosis within a cell population of high cell confluence, or high cell density, which has proven challenging because of the difficulty in separating individual cells. We first detect the candidates for birth events that are defined as the time and location at which mitosis is complete and two daughter cells first appear. Each candidate is then examined whether it is real or not after incorporating spatio-temporal information by tracking the candidate in the neighboring frames. For the examination, we design a probabilistic model named Two-Labeled Hidden Conditional Random Field (TL-HCRF) that can use the information on the timing of the candidate birth event in addition to the visual change of cells over time. Applied to two cell populations of high cell confluence, our method considerably outperforms previous methods. Comparisons with related statistical models also show the superiority of TL-HCRF on the proposed task.

Classification of 3D Multicellular Organization in Phase Microscopy for High Throughput Screening of Therapeutic Targets

“…From our perspective, the main trends in phase contrast microscopy image analysis are (i) cell segmentation and tracking [11,12]; and (ii) cellular event (e.g., mitosis, apoptosis, and cell division) detection [13–15]. However, to the best of our knowledge, the classification of 3D multicellular organization using phase imaging has not been explored, which can actually play an important role for high throughput screening of therapeutic targets.…”

Apoptosis Detection for Adherent Cell Populations in Time-Lapse Phase-Contrast Microscopy Images

“…We tested several other classifiers used for mitosis detection, Hidden Conditional Random Field (HCRF) [6] and its variations [2,7]. All these classifiers as well as an RBF kernel SVM did not outperform a linear SVM despite their higher computational cost, presumably because visual features of apoptosis are less informative and more noisy in the sense that apoptosis does not involve distinctive morphological features, such as a figure eight shape during mitosis.…”

“…Under such a circumstance, a max-margin classifier with a simple decision boundary might be more effective to eliminate outliers or meaningless patterns. After classification, the post-processing in [7] is conducted to prevent one apoptosis from being detected multiple times.…”

Unsupervised Two-Path Neural Network for Cell Event Detection and Classification Using Spatiotemporal Patterns

“…The most important cell events, such as cell death, cell division, and cell differentiation, significantly change the properties of a whole cell. For example, during cell division, a cell shrinks and becomes circular [11], while during cell death, a cell's membrane swells into spherical bubbles [12]. These events may often appear random in nature and are governed by many stochastic elements in biological systems [13,14].…”

“…There have been a number of methods for cell event detection that have been reported over the past 8 years. These include the Hidden Conditional Random Fields [6,11,12,[16][17][18], the semi-Markov [19], autoregression and multi-output Gaussian process methods [20,21]. All these methods required labelled training data and thus they cannot be widely applied to various types of cell events.…”