Limits...
An automated three-dimensional detection and segmentation method for touching cells by integrating concave points clustering and random walker algorithm.

He Y, Meng Y, Gong H, Chen S, Zhang B, Ding W, Luo Q, Li A - PLoS ONE (2014)

Bottom Line: Characterizing cytoarchitecture is crucial for understanding brain functions and neural diseases.In neuroanatomy, it is an important task to accurately extract cell populations' centroids and contours.However, it is difficult to precisely segment numerous cells, especially those cells touching each other.

View Article: PubMed Central - PubMed

Affiliation: Britton Chance Center for Biomedical Photonics, Huazhong University of Science and Technology-Wuhan National Laboratory for Optoelectronics, Wuhan, Hubei, China; MoE Key Laboratory for Biomedical Photonics, Department of Biomedical Engineering, Huazhong University of Science and Technology, Wuhan, Hubei, China.

ABSTRACT
Characterizing cytoarchitecture is crucial for understanding brain functions and neural diseases. In neuroanatomy, it is an important task to accurately extract cell populations' centroids and contours. Recent advances have permitted imaging at single cell resolution for an entire mouse brain using the Nissl staining method. However, it is difficult to precisely segment numerous cells, especially those cells touching each other. As presented herein, we have developed an automated three-dimensional detection and segmentation method applied to the Nissl staining data, with the following two key steps: 1) concave points clustering to determine the seed points of touching cells; and 2) random walker segmentation to obtain cell contours. Also, we have evaluated the performance of our proposed method with several mouse brain datasets, which were captured with the micro-optical sectioning tomography imaging system, and the datasets include closely touching cells. Comparing with traditional detection and segmentation methods, our approach shows promising detection accuracy and high robustness.

Show MeSH

Related in: MedlinePlus

Noise influences on the precision and recall for the K1 stack.Salt and pepper noise is used here. The red point is the detected point, and the green dashed circle indicates a missing seed point caused by noise. (A–D) The volume-rendered (with the colormap's alpha values of 0.5) binary image stack with different levels of noise density, i.e., 0, 0.03, 0.06, and 0.09. (E–H) The seed point detection results with different levels of noise density, i.e., 0, 0.03, 0.06, and 0.09. For easy observation, the seed point is placed on the binary stack, pre-noise elimination, and the pre-noise elimination binary stack is volume-rendered with the color map's alpha values of 0.2. (I) Performance curves of the seed point detection estimation with varying noise densities.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4128780&req=5

pone-0104437-g009: Noise influences on the precision and recall for the K1 stack.Salt and pepper noise is used here. The red point is the detected point, and the green dashed circle indicates a missing seed point caused by noise. (A–D) The volume-rendered (with the colormap's alpha values of 0.5) binary image stack with different levels of noise density, i.e., 0, 0.03, 0.06, and 0.09. (E–H) The seed point detection results with different levels of noise density, i.e., 0, 0.03, 0.06, and 0.09. For easy observation, the seed point is placed on the binary stack, pre-noise elimination, and the pre-noise elimination binary stack is volume-rendered with the color map's alpha values of 0.2. (I) Performance curves of the seed point detection estimation with varying noise densities.

Mentions: To assess the general capabilities of our proposed method, we added noise into the stack and evaluated the seed point detection results. Because the seed point detection and segmentation steps were both performed on the binary image, we chose salt and pepper noise with noise density ranging from 0.00 to 0.10, i.e., {0, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0, 0.07, 0.08, 0.09, 0.10}, and the parameters of the original binary image stack and the noised binary image stack were the same. Finally, we computed the recall and precision values at each noise level. The seed point detection results of the K1 stack with some noise levels are shown in Figure 9A–E. We found that the detection results were influenced by noise to a certain extent. Some seed points were missing (see the green dashed circle in Figure 9F–H) when noise was added. Figure 9I shows the performance curves of the seed point detection estimation (Precision, Recall) for varying noise densities. The precision performance of the seed point detection varied with the added noise, but it never varied by more than 10%, and the recall performance was reduced with an increase in the noise density. In the Ckernel-DBSCAN algorithm, we needed to compute the density of concave points, and the addition of salt and pepper noise can result in false positives for concave points. Thus, the density of concave points was influenced, which may affect the cluster concave point results, for example, the high density of noise points caused concave points of different touching-cell-pairs to be clustered together, reducing the number of clusters and omitting the seed points of some touching cells (recall performance was reduced).


An automated three-dimensional detection and segmentation method for touching cells by integrating concave points clustering and random walker algorithm.

He Y, Meng Y, Gong H, Chen S, Zhang B, Ding W, Luo Q, Li A - PLoS ONE (2014)

Noise influences on the precision and recall for the K1 stack.Salt and pepper noise is used here. The red point is the detected point, and the green dashed circle indicates a missing seed point caused by noise. (A–D) The volume-rendered (with the colormap's alpha values of 0.5) binary image stack with different levels of noise density, i.e., 0, 0.03, 0.06, and 0.09. (E–H) The seed point detection results with different levels of noise density, i.e., 0, 0.03, 0.06, and 0.09. For easy observation, the seed point is placed on the binary stack, pre-noise elimination, and the pre-noise elimination binary stack is volume-rendered with the color map's alpha values of 0.2. (I) Performance curves of the seed point detection estimation with varying noise densities.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4128780&req=5

pone-0104437-g009: Noise influences on the precision and recall for the K1 stack.Salt and pepper noise is used here. The red point is the detected point, and the green dashed circle indicates a missing seed point caused by noise. (A–D) The volume-rendered (with the colormap's alpha values of 0.5) binary image stack with different levels of noise density, i.e., 0, 0.03, 0.06, and 0.09. (E–H) The seed point detection results with different levels of noise density, i.e., 0, 0.03, 0.06, and 0.09. For easy observation, the seed point is placed on the binary stack, pre-noise elimination, and the pre-noise elimination binary stack is volume-rendered with the color map's alpha values of 0.2. (I) Performance curves of the seed point detection estimation with varying noise densities.
Mentions: To assess the general capabilities of our proposed method, we added noise into the stack and evaluated the seed point detection results. Because the seed point detection and segmentation steps were both performed on the binary image, we chose salt and pepper noise with noise density ranging from 0.00 to 0.10, i.e., {0, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0, 0.07, 0.08, 0.09, 0.10}, and the parameters of the original binary image stack and the noised binary image stack were the same. Finally, we computed the recall and precision values at each noise level. The seed point detection results of the K1 stack with some noise levels are shown in Figure 9A–E. We found that the detection results were influenced by noise to a certain extent. Some seed points were missing (see the green dashed circle in Figure 9F–H) when noise was added. Figure 9I shows the performance curves of the seed point detection estimation (Precision, Recall) for varying noise densities. The precision performance of the seed point detection varied with the added noise, but it never varied by more than 10%, and the recall performance was reduced with an increase in the noise density. In the Ckernel-DBSCAN algorithm, we needed to compute the density of concave points, and the addition of salt and pepper noise can result in false positives for concave points. Thus, the density of concave points was influenced, which may affect the cluster concave point results, for example, the high density of noise points caused concave points of different touching-cell-pairs to be clustered together, reducing the number of clusters and omitting the seed points of some touching cells (recall performance was reduced).

Bottom Line: Characterizing cytoarchitecture is crucial for understanding brain functions and neural diseases.In neuroanatomy, it is an important task to accurately extract cell populations' centroids and contours.However, it is difficult to precisely segment numerous cells, especially those cells touching each other.

View Article: PubMed Central - PubMed

Affiliation: Britton Chance Center for Biomedical Photonics, Huazhong University of Science and Technology-Wuhan National Laboratory for Optoelectronics, Wuhan, Hubei, China; MoE Key Laboratory for Biomedical Photonics, Department of Biomedical Engineering, Huazhong University of Science and Technology, Wuhan, Hubei, China.

ABSTRACT
Characterizing cytoarchitecture is crucial for understanding brain functions and neural diseases. In neuroanatomy, it is an important task to accurately extract cell populations' centroids and contours. Recent advances have permitted imaging at single cell resolution for an entire mouse brain using the Nissl staining method. However, it is difficult to precisely segment numerous cells, especially those cells touching each other. As presented herein, we have developed an automated three-dimensional detection and segmentation method applied to the Nissl staining data, with the following two key steps: 1) concave points clustering to determine the seed points of touching cells; and 2) random walker segmentation to obtain cell contours. Also, we have evaluated the performance of our proposed method with several mouse brain datasets, which were captured with the micro-optical sectioning tomography imaging system, and the datasets include closely touching cells. Comparing with traditional detection and segmentation methods, our approach shows promising detection accuracy and high robustness.

Show MeSH
Related in: MedlinePlus