Limits...
A probabilistic cell model in background corrected image sequences for single cell analysis.

Kachouie NN, Fieguth P, Jervis E - Biomed Eng Online (2010)

Bottom Line: The resulting background-removed images have fewer artifacts and allow cells to be localized and detected more reliably.The experimental results generated by applying the proposed method to different Hematopoietic Stem Cell (HSC) image sequences are quite promising.This method can be potentially used for single cell analysis to study the temporal dynamics of cells.

View Article: PubMed Central - HTML - PubMed

Affiliation: Department of Systems Design Engineering, University of Waterloo, Waterloo, Canada. nnezamod@mit.edu.

ABSTRACT

Background: Methods of manual cell localization and outlining are so onerous that automated tracking methods would seem mandatory for handling huge image sequences, nevertheless manual tracking is, astonishingly, still widely practiced in areas such as cell biology which are outside the influence of most image processing research. The goal of our research is to address this gap by developing automated methods of cell tracking, localization, and segmentation. Since even an optimal frame-to-frame association method cannot compensate and recover from poor detection, it is clear that the quality of cell tracking depends on the quality of cell detection within each frame.

Methods: Cell detection performs poorly where the background is not uniform and includes temporal illumination variations, spatial non-uniformities, and stationary objects such as well boundaries (which confine the cells under study). To improve cell detection, the signal to noise ratio of the input image can be increased via accurate background estimation. In this paper we investigate background estimation, for the purpose of cell detection. We propose a cell model and a method for background estimation, driven by the proposed cell model, such that well structure can be identified, and explicitly rejected, when estimating the background.

Results: The resulting background-removed images have fewer artifacts and allow cells to be localized and detected more reliably. The experimental results generated by applying the proposed method to different Hematopoietic Stem Cell (HSC) image sequences are quite promising.

Conclusion: The understanding of cell behavior relies on precise information about the temporal dynamics and spatial distribution of cells. Such information may play a key role in disease research and regenerative medicine, so automated methods for observation and measurement of cells from microscopic images are in high demand. The proposed method in this paper is capable of localizing single cells in microwells and can be adapted for the other cell types that may not have circular shape. This method can be potentially used for single cell analysis to study the temporal dynamics of cells.

Show MeSH
Background estimation and cell detection. (a) The proposed cell model: (i) Coarsely cropped well in which well boundaries are visible. (ii) Applying the cell model to (i). In contrast with the perfect result obtained by applying the cell model to a cropped well interior with no boundaries, here the cell model performs very poorly where the cropped well contains visible well boundaries. (iii) Corrected well image after point-wise background subtraction. (iv) Applying the cell model to (iii). (b) The process of the proposed cell model and background estimation method: (i) Point-wise estimated background (). (ii) Subtracting out the estimated background image obtained in (i) from the original image. (iii) Located cell centers applying cell model P in (8).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC2967554&req=5

Figure 3: Background estimation and cell detection. (a) The proposed cell model: (i) Coarsely cropped well in which well boundaries are visible. (ii) Applying the cell model to (i). In contrast with the perfect result obtained by applying the cell model to a cropped well interior with no boundaries, here the cell model performs very poorly where the cropped well contains visible well boundaries. (iii) Corrected well image after point-wise background subtraction. (iv) Applying the cell model to (iii). (b) The process of the proposed cell model and background estimation method: (i) Point-wise estimated background (). (ii) Subtracting out the estimated background image obtained in (i) from the original image. (iii) Located cell centers applying cell model P in (8).

Mentions: The threshold is computed analogously to [8], in which the threshold is varied and selected to minimize the sum of missed detections and false alarms. For the proposed cell metric (8), a threshold of 0.25 was found to be very effective, giving a detection rate over 95%, and a false alarm rate of approximately 6%. Fig. 2(b) shows the application of the cell model to a cropped well interior with no boundaries (Fig. 2(b-i)) and a coarsely cropped well before and after background correction is depicted in Figs. 3(a-i) and 3(a-iii) respectively. As it can be observed, the cell model performs very poorly (Fig. 3(a-ii)) where the cropped well contains visible well boundaries, however the cell model performs perfectly where it is applied to a cropped well interior with no boundaries (Fig. 2(b-ii)) or a background corrected coarsely cropped well (Fig. 3(a-iv)).


A probabilistic cell model in background corrected image sequences for single cell analysis.

Kachouie NN, Fieguth P, Jervis E - Biomed Eng Online (2010)

Background estimation and cell detection. (a) The proposed cell model: (i) Coarsely cropped well in which well boundaries are visible. (ii) Applying the cell model to (i). In contrast with the perfect result obtained by applying the cell model to a cropped well interior with no boundaries, here the cell model performs very poorly where the cropped well contains visible well boundaries. (iii) Corrected well image after point-wise background subtraction. (iv) Applying the cell model to (iii). (b) The process of the proposed cell model and background estimation method: (i) Point-wise estimated background (). (ii) Subtracting out the estimated background image obtained in (i) from the original image. (iii) Located cell centers applying cell model P in (8).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC2967554&req=5

Figure 3: Background estimation and cell detection. (a) The proposed cell model: (i) Coarsely cropped well in which well boundaries are visible. (ii) Applying the cell model to (i). In contrast with the perfect result obtained by applying the cell model to a cropped well interior with no boundaries, here the cell model performs very poorly where the cropped well contains visible well boundaries. (iii) Corrected well image after point-wise background subtraction. (iv) Applying the cell model to (iii). (b) The process of the proposed cell model and background estimation method: (i) Point-wise estimated background (). (ii) Subtracting out the estimated background image obtained in (i) from the original image. (iii) Located cell centers applying cell model P in (8).
Mentions: The threshold is computed analogously to [8], in which the threshold is varied and selected to minimize the sum of missed detections and false alarms. For the proposed cell metric (8), a threshold of 0.25 was found to be very effective, giving a detection rate over 95%, and a false alarm rate of approximately 6%. Fig. 2(b) shows the application of the cell model to a cropped well interior with no boundaries (Fig. 2(b-i)) and a coarsely cropped well before and after background correction is depicted in Figs. 3(a-i) and 3(a-iii) respectively. As it can be observed, the cell model performs very poorly (Fig. 3(a-ii)) where the cropped well contains visible well boundaries, however the cell model performs perfectly where it is applied to a cropped well interior with no boundaries (Fig. 2(b-ii)) or a background corrected coarsely cropped well (Fig. 3(a-iv)).

Bottom Line: The resulting background-removed images have fewer artifacts and allow cells to be localized and detected more reliably.The experimental results generated by applying the proposed method to different Hematopoietic Stem Cell (HSC) image sequences are quite promising.This method can be potentially used for single cell analysis to study the temporal dynamics of cells.

View Article: PubMed Central - HTML - PubMed

Affiliation: Department of Systems Design Engineering, University of Waterloo, Waterloo, Canada. nnezamod@mit.edu.

ABSTRACT

Background: Methods of manual cell localization and outlining are so onerous that automated tracking methods would seem mandatory for handling huge image sequences, nevertheless manual tracking is, astonishingly, still widely practiced in areas such as cell biology which are outside the influence of most image processing research. The goal of our research is to address this gap by developing automated methods of cell tracking, localization, and segmentation. Since even an optimal frame-to-frame association method cannot compensate and recover from poor detection, it is clear that the quality of cell tracking depends on the quality of cell detection within each frame.

Methods: Cell detection performs poorly where the background is not uniform and includes temporal illumination variations, spatial non-uniformities, and stationary objects such as well boundaries (which confine the cells under study). To improve cell detection, the signal to noise ratio of the input image can be increased via accurate background estimation. In this paper we investigate background estimation, for the purpose of cell detection. We propose a cell model and a method for background estimation, driven by the proposed cell model, such that well structure can be identified, and explicitly rejected, when estimating the background.

Results: The resulting background-removed images have fewer artifacts and allow cells to be localized and detected more reliably. The experimental results generated by applying the proposed method to different Hematopoietic Stem Cell (HSC) image sequences are quite promising.

Conclusion: The understanding of cell behavior relies on precise information about the temporal dynamics and spatial distribution of cells. Such information may play a key role in disease research and regenerative medicine, so automated methods for observation and measurement of cells from microscopic images are in high demand. The proposed method in this paper is capable of localizing single cells in microwells and can be adapted for the other cell types that may not have circular shape. This method can be potentially used for single cell analysis to study the temporal dynamics of cells.

Show MeSH