WaterBalloons: A hybrid watershed Balloon Snake segmentation

Dagher, Issam; Tom, K. El

doi:10.1016/j.imavis.2007.10.010

Cited by 45 publications

(16 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Their aim is to delineate cell borders using such techniques as: local greyscale thresholding followed by scissoring and morphological thinning [4], [5], scale-space filtering followed by binarization and morphological processing [6] or hexagon detection using shape dependent filters [7], [8], [9]. More sophisticated methods include application of watersheds [10], [11], [12], [13], [14], active contours [15], [16], genetic algorithms [17] or analysis of local pixel levels aimed at finding intensity valleys corresponding to borders between cells [18]. Several machine learning approaches have also been proposed by the team of Ruggeri, including: neural network [19], [20], Bayesian framework [21], support vector machines classifier [22] and genetic algorithm [23].…”

Section: Introductionmentioning

confidence: 99%

Corneal Endothelium Image Segmentation Using Feedforward Neural Network

Fabijańska

2017

Proceedings of the 2017 Federated Conference on Computer Science and Information Systems

View full text Add to dashboard Cite

Abstract-In this paper the problem of corneal endothelium image segmentation is considered. Particularly, a fully automatic approach for delineating contours of corneal endothelial cells is proposed. The approach produces one pixel width outline of cells. It bases on a simple feedforward neural network trained to recognize pixels which belong to the cell borders. The edge probability (edginess) map output by the network is next analysed row by row and column by column in order to find local peaks of the network response. These peaks are considered as cell border candidates and in the last step of the method via binary morphological processing are linked to create continuous outlines of cells. The results of applying the proposed approach to publicity available data set of corneal endothelium images as well as the assessment of the method against ground truth segmentation are presented and discussed. Obtained results show, that the proposed approach performs very well. The resulting mean absolute error of cell number determination is around 5% while the average DICE measure reaches 0.83 which is a good result, especially when one pixel width objects are compared.

show abstract

Section: Introductionmentioning

confidence: 99%

Corneal Endothelium Image Segmentation Using Feedforward Neural Network

Fabijańska

2017

Proceedings of the 2017 Federated Conference on Computer Science and Information Systems

View full text Add to dashboard Cite

show abstract

“…Tsai et al [8] replaced the thresholding step with k-means clustering into two partitions. Dagher and Tom [9] combined the watershed segmentation algorithm with the active contour model by using the watershed segmentation result of a down-sampled image as the initial contour of the snake for the segmentation of blood cells and corneal cells. Huang and Lai [10] also used the marker-based watershed segmentation algorithm to find an approximate segmentation, applied heuristic rules to eliminate non-nuclei regions, and used active contours to improve the nuclei boundaries in biopsy images of liver cells.…”

Section: Introductionmentioning

confidence: 99%

“…Most of these methods focus on the segmentation of only the nuclei [5,6,9,10,12,13] for which there is relatively higher contrast around the boundaries. However, detection of the cytoplasm regions is also crucial because cytoplasm features have been shown to be very useful for the identification of abnormal cells [15].…”

Section: Introductionmentioning

confidence: 99%

Unsupervised segmentation and classification of cervical cell images

Genctav

Aksoy

Önder

2012

Pattern Recognition

226

176

View full text Add to dashboard Cite

a b s t r a c tThe Pap smear test is a manual screening procedure that is used to detect precancerous changes in cervical cells based on color and shape properties of their nuclei and cytoplasms. Automating this procedure is still an open problem due to the complexities of cell structures. In this paper, we propose an unsupervised approach for the segmentation and classification of cervical cells. The segmentation process involves automatic thresholding to separate the cell regions from the background, a multi-scale hierarchical segmentation algorithm to partition these regions based on homogeneity and circularity, and a binary classifier to finalize the separation of nuclei from cytoplasm within the cell regions. Classification is posed as a grouping problem by ranking the cells based on their feature characteristics modeling abnormality degrees. The proposed procedure constructs a tree using hierarchical clustering, and then arranges the cells in a linear order by using an optimal leaf ordering algorithm that maximizes the similarity of adjacent leaves without any requirement for training examples or parameter adjustment. Performance evaluation using two data sets show the effectiveness of the proposed approach in images having inconsistent staining, poor contrast, and overlapping cells.

show abstract

“…Bernander et al (2013); Malmberg et al (2014), and Selig et al (2015a) developed fast robust stochastic watershed algorithms based on a probability density function that can locate silent contours without wrong matches. There is also a group of more sophisticated approaches: applying Bayesian framework supported by simulated annealing for cell border location (Foracchia and Ruggeri, 2003) improved by statistical description (Foracchia and Ruggeri, 2007); exploiting active contours (Dagher and El Tom, 2008), snake-lets (Charłampowicz et al, 2014), level sets (Zhou, 2007), and wavelets (Khan et al, 2007). The data mining and rough sets theory found an application in the solution designed by Poletti and Ruggeri (2014); while Ruggeri et al (2010) used an artificial neural network to classify whether a pixel belongs to a cell body or boundary oriented at different angles; the genetic algorithm was exploited by Ruggeri and Scarpa (2015); Scarpa and Ruggeri (2015), who supported it with information about pixel intensities and regularity of cell shapes.…”

Section: Determination Of Cell Locationmentioning

confidence: 99%

The Correlation Analysis of the Shape Parameters for Endothelial Image Characterisation

Nurzynska

Piórkowski

2016

Image Anal Stereol

View full text Add to dashboard Cite

Microscopic images of corneal endothelium cells are investigated to deliver information about their medical state. Although this could be achieved automatically, this examination is manual and very time consuming. Two medical parameters for endothelial layer quality description have been introduced and more are planned. Yet, since they will exploit image processing, a thoughtful overview of applicable existing shape parameters is necessary. This work investigates the possibility of exploiting well-known image processing techniques for describing the endothelial layer by calculating information about shape features using spatial moments or topological attributes. The comparison concentrates on finding which shape measures could be combined to improve descriptions, and which cannot due to their high correlation and the fact that they do not contain any new information. The performed experiments revealed a set of 17 non-correlated features and four groups of shape parameters that show some correlation, but one representative can always be selected. Moreover, the investigation proved some correlation between the metrics used in medicine and considered shape features.

show abstract

WaterBalloons: A hybrid watershed Balloon Snake segmentation

Cited by 45 publications

References 14 publications

Corneal Endothelium Image Segmentation Using Feedforward Neural Network

Corneal Endothelium Image Segmentation Using Feedforward Neural Network

Unsupervised segmentation and classification of cervical cell images

The Correlation Analysis of the Shape Parameters for Endothelial Image Characterisation

Contact Info

Product

Resources

About