Limits...
Segmental HOG: new descriptor for glomerulus detection in kidney microscopy image.

Kato T, Relator R, Ngouv H, Hirohashi Y, Takaki O, Kakimoto T, Okada K - BMC Bioinformatics (2015)

Bottom Line: Moreover, the novel segmentation technique employed herewith generates high-quality segmentation outputs, and the algorithm is assured to converge to an optimal solution.Consequently, experiments using real-world image data revealed that Segmental HOG achieved significant improvements in detection performance compared to Rectangular HOG.The proposed descriptor for glomeruli detection presents promising results, and it is expected to be useful in pathological evaluation.

View Article: PubMed Central - PubMed

Affiliation: Faculty of Science and Engineering, Gunma University, Kiryu-shi, Gunma, 376-8515, Japan. katotsu@cs.gunma-u.ac.jp.

ABSTRACT

Background: The detection of the glomeruli is a key step in the histopathological evaluation of microscopic images of the kidneys. However, the task of automatic detection of the glomeruli poses challenges owing to the differences in their sizes and shapes in renal sections as well as the extensive variations in their intensities due to heterogeneity in immunohistochemistry staining. Although the rectangular histogram of oriented gradients (Rectangular HOG) is a widely recognized powerful descriptor for general object detection, it shows many false positives owing to the aforementioned difficulties in the context of glomeruli detection.

Results: A new descriptor referred to as Segmental HOG was developed to perform a comprehensive detection of hundreds of glomeruli in images of whole kidney sections. The new descriptor possesses flexible blocks that can be adaptively fitted to input images in order to acquire robustness for the detection of the glomeruli. Moreover, the novel segmentation technique employed herewith generates high-quality segmentation outputs, and the algorithm is assured to converge to an optimal solution. Consequently, experiments using real-world image data revealed that Segmental HOG achieved significant improvements in detection performance compared to Rectangular HOG.

Conclusion: The proposed descriptor for glomeruli detection presents promising results, and it is expected to be useful in pathological evaluation.

No MeSH data available.


Runtime comparisons. In Panel (a), ndp of DCDP with three splitting schemes and EDP is shown, where ndp is the number of invoking the O(mnς) DP routine. The number of iterates of MPLP and MPLP+ is plotted in Panel (b), where the time complexity for one iterate in MPLP and MPLP+ is O(mnς), which is equal to one DP routine. The computational time of each algorithm is plotted in Panels (c) and (d)
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC4590714&req=5

Fig9: Runtime comparisons. In Panel (a), ndp of DCDP with three splitting schemes and EDP is shown, where ndp is the number of invoking the O(mnς) DP routine. The number of iterates of MPLP and MPLP+ is plotted in Panel (b), where the time complexity for one iterate in MPLP and MPLP+ is O(mnς), which is equal to one DP routine. The computational time of each algorithm is plotted in Panels (c) and (d)

Mentions: The computational time of the new segmentation algorithm, DCDP, is compared with that of EDP. The two algorithms solve the same optimization problem, and both algorithms always find the same optimal solution. DCDP and EDP are implemented in C++ language, and the runtimes are measured on a Linux machine with Intel(R) Core(TM) i7 CPU and 8-GB memory. First, the number of times when the O(nmς) DP routine was invoked, which we denote by ndp, is counted using the annotated glomeruli in Set B. Figure 9a shows the box-plot of ndp for all methods. While the value of ndp for EDP is always n, the values for DCDP depend on the input images and the splitting schemes, Half Split, Max Split, and Adap Split. For 46.32 % of glomeruli, the optimal solutions are found within the first DP routine (i.e. ndp=1). The medians of the ndp’s when using Half Split, Max Split, and Adap Split are 5, 3, and 3, respectively. In other words, the medians of the depth of the branching tree for each scheme are 3, 2, and 2, respectively, and the respective 75th percentiles of ndp’s are 11, 7, and 5. For Adap Split, there is no case where ndp is larger than n, whereas the number of glomeruli with ndp>n are 4 (0.40 %) and 16 (1.61 %) for Half Split and Max Split, respectively. This implies that Adap Split is the best heuristic process among the three splitting schemes.Fig. 9


Segmental HOG: new descriptor for glomerulus detection in kidney microscopy image.

Kato T, Relator R, Ngouv H, Hirohashi Y, Takaki O, Kakimoto T, Okada K - BMC Bioinformatics (2015)

Runtime comparisons. In Panel (a), ndp of DCDP with three splitting schemes and EDP is shown, where ndp is the number of invoking the O(mnς) DP routine. The number of iterates of MPLP and MPLP+ is plotted in Panel (b), where the time complexity for one iterate in MPLP and MPLP+ is O(mnς), which is equal to one DP routine. The computational time of each algorithm is plotted in Panels (c) and (d)
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC4590714&req=5

Fig9: Runtime comparisons. In Panel (a), ndp of DCDP with three splitting schemes and EDP is shown, where ndp is the number of invoking the O(mnς) DP routine. The number of iterates of MPLP and MPLP+ is plotted in Panel (b), where the time complexity for one iterate in MPLP and MPLP+ is O(mnς), which is equal to one DP routine. The computational time of each algorithm is plotted in Panels (c) and (d)
Mentions: The computational time of the new segmentation algorithm, DCDP, is compared with that of EDP. The two algorithms solve the same optimization problem, and both algorithms always find the same optimal solution. DCDP and EDP are implemented in C++ language, and the runtimes are measured on a Linux machine with Intel(R) Core(TM) i7 CPU and 8-GB memory. First, the number of times when the O(nmς) DP routine was invoked, which we denote by ndp, is counted using the annotated glomeruli in Set B. Figure 9a shows the box-plot of ndp for all methods. While the value of ndp for EDP is always n, the values for DCDP depend on the input images and the splitting schemes, Half Split, Max Split, and Adap Split. For 46.32 % of glomeruli, the optimal solutions are found within the first DP routine (i.e. ndp=1). The medians of the ndp’s when using Half Split, Max Split, and Adap Split are 5, 3, and 3, respectively. In other words, the medians of the depth of the branching tree for each scheme are 3, 2, and 2, respectively, and the respective 75th percentiles of ndp’s are 11, 7, and 5. For Adap Split, there is no case where ndp is larger than n, whereas the number of glomeruli with ndp>n are 4 (0.40 %) and 16 (1.61 %) for Half Split and Max Split, respectively. This implies that Adap Split is the best heuristic process among the three splitting schemes.Fig. 9

Bottom Line: Moreover, the novel segmentation technique employed herewith generates high-quality segmentation outputs, and the algorithm is assured to converge to an optimal solution.Consequently, experiments using real-world image data revealed that Segmental HOG achieved significant improvements in detection performance compared to Rectangular HOG.The proposed descriptor for glomeruli detection presents promising results, and it is expected to be useful in pathological evaluation.

View Article: PubMed Central - PubMed

Affiliation: Faculty of Science and Engineering, Gunma University, Kiryu-shi, Gunma, 376-8515, Japan. katotsu@cs.gunma-u.ac.jp.

ABSTRACT

Background: The detection of the glomeruli is a key step in the histopathological evaluation of microscopic images of the kidneys. However, the task of automatic detection of the glomeruli poses challenges owing to the differences in their sizes and shapes in renal sections as well as the extensive variations in their intensities due to heterogeneity in immunohistochemistry staining. Although the rectangular histogram of oriented gradients (Rectangular HOG) is a widely recognized powerful descriptor for general object detection, it shows many false positives owing to the aforementioned difficulties in the context of glomeruli detection.

Results: A new descriptor referred to as Segmental HOG was developed to perform a comprehensive detection of hundreds of glomeruli in images of whole kidney sections. The new descriptor possesses flexible blocks that can be adaptively fitted to input images in order to acquire robustness for the detection of the glomeruli. Moreover, the novel segmentation technique employed herewith generates high-quality segmentation outputs, and the algorithm is assured to converge to an optimal solution. Consequently, experiments using real-world image data revealed that Segmental HOG achieved significant improvements in detection performance compared to Rectangular HOG.

Conclusion: The proposed descriptor for glomeruli detection presents promising results, and it is expected to be useful in pathological evaluation.

No MeSH data available.