Cell Segmentation for Quantitative Analysis of Anodized TiO$_2$ Foil

Lee, Sang Chul

doi:10.1109/tii.2018.2866804

Cited by 6 publications

(4 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To segment an image with resolution 200×200 into two classes, the total number of outputs is 2 40000 , which needs infinite training data. The slope difference distribution based threshold selection method performed better than other image segmentation methods for cell segmentation [7], [8]. However, it did not achieve the perfect performance yet because of the significantly varied intensities of cells in most cell images.…”

Section: B Discussionmentioning

confidence: 92%

“…where T 1 is the optimal intensity threshold calculated by slope difference distribution (SDD) [8]. T 2 is the low intensity threshold to segment the cells with lower intensities and T 3 is the high intensity threshold to segment the cells with higher intensities.…”

Section: The Proposed Cell Detection Methodsmentioning

confidence: 99%

“…S 4 is then filtered by the noise blob removing filter [8] to generate the segmentation S 4 for the cells with middle intensities. The fusion segmentation is computed as:…”

Section: The Proposed Cell Detection Methodsmentioning

confidence: 99%

“…w is a weighting coefficient and its default value is 0.5. With the area threshold T 5 , the seeds S seed are computed with the iterative erosion [8]. T 6 is the area threshold to remove fake seeds and it is determined based on the priori knowledge of the specific type of cells to be segmented.…”

Section: The Proposed Cell Detection Methodsmentioning

confidence: 99%

See 3 more Smart Citations

A New Approach for Cell Detection and Tracking

2019

View full text Add to dashboard Cite

In recent years, cell tracking methods by detection have become more and more popular because they outperformed cell tracking methods by contour evolution in most practical cell tracking applications. Yet, the most frequently used segmentation technique by cell detection methods is still threshold selection that is determined manually or by algorithms proposed in the 1970s. As a whole, these old threshold selection methods could not meet the accuracy requirement of cell detection adequately. In this paper, we propose a new approach of cell tracking by detection based on a multiple-threshold segmentation method that calculates multiple thresholds automatically and robustly. After cell detection, the proposed approach generates the timeline moving trajectory of a cell by connecting the cell positions along the time lapse image sequences based on morphological operations. We use four types of cells to verify the effectiveness of the proposed approach and the experimental results are favorable.

show abstract

Section: B Discussionmentioning

confidence: 92%

Section: The Proposed Cell Detection Methodsmentioning

confidence: 99%

“…S 4 is then filtered by the noise blob removing filter [8] to generate the segmentation S 4 for the cells with middle intensities. The fusion segmentation is computed as:…”

Section: The Proposed Cell Detection Methodsmentioning

confidence: 99%

Section: The Proposed Cell Detection Methodsmentioning

confidence: 99%

See 2 more Smart Citations