Limits...
Method: automatic segmentation of mitochondria utilizing patch classification, contour pair classification, and automatically seeded level sets.

Giuly RJ, Martone ME, Ellisman MH - BMC Bioinformatics (2012)

Bottom Line: We show that our final level set operation, which is automatically seeded with output from previous steps, helps to smooth the results.Overall, our results show that use of contour pair classification and level set operations improve segmentation accuracy beyond patch classification alone.We show that the Cytoseg process performs well compared to another modern technique based on Radon-Like Features.

View Article: PubMed Central - HTML - PubMed

Affiliation: Center for Research in Biological Systems, University of California, 9500 Gilman Dr., La Jolla, CA 92093, USA. rgiuly@gmail.com

ABSTRACT

Background: While progress has been made to develop automatic segmentation techniques for mitochondria, there remains a need for more accurate and robust techniques to delineate mitochondria in serial blockface scanning electron microscopic data. Previously developed texture based methods are limited for solving this problem because texture alone is often not sufficient to identify mitochondria. This paper presents a new three-step method, the Cytoseg process, for automated segmentation of mitochondria contained in 3D electron microscopic volumes generated through serial block face scanning electron microscopic imaging. The method consists of three steps. The first is a random forest patch classification step operating directly on 2D image patches. The second step consists of contour-pair classification. At the final step, we introduce a method to automatically seed a level set operation with output from previous steps.

Results: We report accuracy of the Cytoseg process on three types of tissue and compare it to a previous method based on Radon-Like Features. At step 1, we show that the patch classifier identifies mitochondria texture but creates many false positive pixels. At step 2, our contour processing step produces contours and then filters them with a second classification step, helping to improve overall accuracy. We show that our final level set operation, which is automatically seeded with output from previous steps, helps to smooth the results. Overall, our results show that use of contour pair classification and level set operations improve segmentation accuracy beyond patch classification alone. We show that the Cytoseg process performs well compared to another modern technique based on Radon-Like Features.

Conclusions: We demonstrated that texture based methods for mitochondria segmentation can be enhanced with multiple steps that form an image processing pipeline. While we used a random-forest based patch classifier to recognize texture, it would be possible to replace this with other texture identifiers, and we plan to explore this in future work.

Show MeSH
Smoothing effect of Step 3. (a) Stack of output contours representing one mitochondria from step 2 using data from Test 1. Contours are rendered as solid slabs. (b) Output of step 3. Notice that the level set operation yields a result that is smoothed. (The slice view for (a) and (b) is 3.5 μm × 3.5 μm in size.)
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3293777&req=5

Figure 6: Smoothing effect of Step 3. (a) Stack of output contours representing one mitochondria from step 2 using data from Test 1. Contours are rendered as solid slabs. (b) Output of step 3. Notice that the level set operation yields a result that is smoothed. (The slice view for (a) and (b) is 3.5 μm × 3.5 μm in size.)

Mentions: In the third step, a "geodesic active contour" level set filter is then used to produce 3D blobs that represent mitochondria. The geodesic active contour filter takes two input images (1) the result of an input fast march filter and (2) the edge potential map. In our method, the input of the fast march filter is the probability map from step 1. For the edge potential map, we used the gradient magnitude of the probability map from step 1. The filter performs a level set operation to generate a final result. Proper seeding of the input fast march is critical to produce an accurate segmentation. The inner region of the contours detected in step 2 demarcates the initial seed points, while the gradient magnitude of the 3D probability map defines the edge potential image. The inner region of contours is found by taking the full region of the contour and eroding it by 10 pixels. The edge potential image produces a "force" so that the geodesic active contour filter operation tends to fill until it reaches the boundary of mitochondria and non-mitochondria voxels. Results are shown in Figure 2(e) and 6(b).


Method: automatic segmentation of mitochondria utilizing patch classification, contour pair classification, and automatically seeded level sets.

Giuly RJ, Martone ME, Ellisman MH - BMC Bioinformatics (2012)

Smoothing effect of Step 3. (a) Stack of output contours representing one mitochondria from step 2 using data from Test 1. Contours are rendered as solid slabs. (b) Output of step 3. Notice that the level set operation yields a result that is smoothed. (The slice view for (a) and (b) is 3.5 μm × 3.5 μm in size.)
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3293777&req=5

Figure 6: Smoothing effect of Step 3. (a) Stack of output contours representing one mitochondria from step 2 using data from Test 1. Contours are rendered as solid slabs. (b) Output of step 3. Notice that the level set operation yields a result that is smoothed. (The slice view for (a) and (b) is 3.5 μm × 3.5 μm in size.)
Mentions: In the third step, a "geodesic active contour" level set filter is then used to produce 3D blobs that represent mitochondria. The geodesic active contour filter takes two input images (1) the result of an input fast march filter and (2) the edge potential map. In our method, the input of the fast march filter is the probability map from step 1. For the edge potential map, we used the gradient magnitude of the probability map from step 1. The filter performs a level set operation to generate a final result. Proper seeding of the input fast march is critical to produce an accurate segmentation. The inner region of the contours detected in step 2 demarcates the initial seed points, while the gradient magnitude of the 3D probability map defines the edge potential image. The inner region of contours is found by taking the full region of the contour and eroding it by 10 pixels. The edge potential image produces a "force" so that the geodesic active contour filter operation tends to fill until it reaches the boundary of mitochondria and non-mitochondria voxels. Results are shown in Figure 2(e) and 6(b).

Bottom Line: We show that our final level set operation, which is automatically seeded with output from previous steps, helps to smooth the results.Overall, our results show that use of contour pair classification and level set operations improve segmentation accuracy beyond patch classification alone.We show that the Cytoseg process performs well compared to another modern technique based on Radon-Like Features.

View Article: PubMed Central - HTML - PubMed

Affiliation: Center for Research in Biological Systems, University of California, 9500 Gilman Dr., La Jolla, CA 92093, USA. rgiuly@gmail.com

ABSTRACT

Background: While progress has been made to develop automatic segmentation techniques for mitochondria, there remains a need for more accurate and robust techniques to delineate mitochondria in serial blockface scanning electron microscopic data. Previously developed texture based methods are limited for solving this problem because texture alone is often not sufficient to identify mitochondria. This paper presents a new three-step method, the Cytoseg process, for automated segmentation of mitochondria contained in 3D electron microscopic volumes generated through serial block face scanning electron microscopic imaging. The method consists of three steps. The first is a random forest patch classification step operating directly on 2D image patches. The second step consists of contour-pair classification. At the final step, we introduce a method to automatically seed a level set operation with output from previous steps.

Results: We report accuracy of the Cytoseg process on three types of tissue and compare it to a previous method based on Radon-Like Features. At step 1, we show that the patch classifier identifies mitochondria texture but creates many false positive pixels. At step 2, our contour processing step produces contours and then filters them with a second classification step, helping to improve overall accuracy. We show that our final level set operation, which is automatically seeded with output from previous steps, helps to smooth the results. Overall, our results show that use of contour pair classification and level set operations improve segmentation accuracy beyond patch classification alone. We show that the Cytoseg process performs well compared to another modern technique based on Radon-Like Features.

Conclusions: We demonstrated that texture based methods for mitochondria segmentation can be enhanced with multiple steps that form an image processing pipeline. While we used a random-forest based patch classifier to recognize texture, it would be possible to replace this with other texture identifiers, and we plan to explore this in future work.

Show MeSH