scispace - formally typeset
Search or ask a question
Journal Article•DOI•

Local optimization based segmentation of spatially-recurring, multi-region objects with part configuration constraints.

Masoud Nosrati1, Ghassan Hamarneh1•
Simon Fraser University1
12 May 2014-IEEE Transactions on Medical Imaging (IEEE)-Vol. 33, Iss: 9, pp 1845-1859
TL;DR: This paper augments the level set framework with the ability to handle these two intuitive geometric relationships, containment and exclusion, along with a distance constraint between boundaries of multi-region objects, and compared this framework with its counterpart methods in the discrete domain.
Abstract: Incorporating prior knowledge into image segmentation algorithms has proven useful for obtaining more accurate and plausible results. Two important constraints, containment and exclusion of regions, have gained attention in recent years mainly due to their descriptive power. In this paper, we augment the level set framework with the ability to handle these two intuitive geometric relationships, containment and exclusion, along with a distance constraint between boundaries of multi-region objects. Level set's important property of automatically handling topological changes of evolving contours/surfaces enables us to segment spatially-recurring objects (e.g., multiple instances of multi-region cells in a large microscopy image) while satisfying the two aforementioned constraints. In addition, the level set approach gives us a very simple and natural way to compute the distance between contours/surfaces and impose constraints on it. The downside, however, is a local optimization framework in which the final segmentation solution depends on the initialization. In fact, here, we sacrifice the optimizability (local instead of global solution) in exchange for lower space complexity (less memory usage) and faster runtime (especially for large microscopic images) as well as no grid artifacts. Nevertheless, the result from validating our method on several biomedical applications showed the utility and advantages of this augmented level set framework (even with rough initialization that is distant from the desired boundaries). We also compared our framework with its counterpart methods in the discrete domain and reported the pros and cons of each of these methods in terms of metrication error and efficiency in memory usage and runtime.

Content maybe subject to copyright    Report

Citations
More filters
Book Chapter•DOI•
Topology Aware Fully Convolutional Networks for Histology Gland Segmentation

[...]

Aïcha BenTaieb1, Ghassan Hamarneh1•
Simon Fraser University1
17 Oct 2016
TL;DR: This work introduces the first deep network trained to encode geometric and topological priors of containment and detachment, and shows how the topology aware architecture outperforms competing methods by up to 10 % in both pixel-level accuracy and object-level Dice.
Abstract: The recent success of deep learning techniques in classification and object detection tasks has been leveraged for segmentation tasks. However, a weakness of these deep segmentation models is their limited ability to encode high level shape priors, such as smoothness and preservation of complex interactions between object regions, which can result in implausible segmentations. In this work, by formulating and optimizing a new loss, we introduce the first deep network trained to encode geometric and topological priors of containment and detachment. Our results on the segmentation of histology glands from a dataset of 165 images demonstrate the advantage of our novel loss terms and show how our topology aware architecture outperforms competing methods by up to 10 % in both pixel-level accuracy and object-level Dice.

170 citations

Journal Article•DOI•
Micro-Net: A unified model for segmentation of various objects in microscopy images.

[...]

Shan E Ahmed Raza1, Shan E Ahmed Raza2, Linda Cheung1, Muhammad Shaban1, Simon Graham1, David Epstein1, Stella Pelengaris1, Michael Khan1, Nasir M. Rajpoot3, Nasir M. Rajpoot1, Nasir M. Rajpoot4 •
University of Warwick1, Institute of Cancer Research2, The Turing Institute3, Coventry Health Care4
01 Feb 2019-Medical Image Analysis
TL;DR: The proposed Micro‐Net is aimed at better object localization in the face of varying intensities and is robust to noise, and compares the results on publicly available data sets and shows that the proposed network outperforms recent deep learning algorithms.

154 citations

Cites background from "Local optimization based segmentati..."

  • ...Nosrati and Hamarneh (2014) and Cohen et al. (2015) first classify tissue regions into different constituents and then employ a constrained level set algorithm to segment the glands. Sirinukunwattana et al. (2015) identified epithelial superpixels and 85 used epithelial regions as vertices of a polygon approximating the boundary of a gland....

    [...]

  • ...Nosrati and Hamarneh (2014) and Cohen et al. (2015) first classify tissue regions into different constituents and then employ a constrained level set algorithm to segment the glands....

    [...]

Journal Article•DOI•
Evaluation of Three Algorithms for the Segmentation of Overlapping Cervical Cells

[...]

Zhi Lu1, Gustavo Carneiro1, Andrew P. Bradley2, Daniela Ushizima3, Masoud Nosrati4, Andrea Gomes Campos Bianchi5, Cláudia Martins Carneiro5, Ghassan Hamarneh4 •
University of Adelaide1, University of Queensland2, Lawrence Berkeley National Laboratory3, Simon Fraser University4, Universidade Federal de Ouro Preto5
01 Mar 2017-IEEE Journal of Biomedical and Health Informatics
TL;DR: The first Overlapping Cervical Cytology Image Segmentation Challenge as discussed by the authors was organized to encourage the development and benchmarking of techniques capable of segmenting individual cells from overlapping cellular clumps in cervical cytology images.
Abstract: In this paper, we introduce and evaluate the systems submitted to the first Overlapping Cervical Cytology Image Segmentation Challenge, held in conjunction with the IEEE International Symposium on Biomedical Imaging 2014. This challenge was organized to encourage the development and benchmarking of techniques capable of segmenting individual cells from overlapping cellular clumps in cervical cytology images, which is a prerequisite for the development of the next generation of computer-aided diagnosis systems for cervical cancer. In particular, these automated systems must detect and accurately segment both the nucleus and cytoplasm of each cell, even when they are clumped together and, hence, partially occluded. However, this is an unsolved problem due to the poor contrast of cytoplasm boundaries, the large variation in size and shape of cells, and the presence of debris and the large degree of cellular overlap. The challenge initially utilized a database of $16$ high-resolution ( $\times$ 40 magnification) images of complex cellular fields of view, in which the isolated real cells were used to construct a database of $945$ cervical cytology images synthesized with a varying number of cells and degree of overlap, in order to provide full access of the segmentation ground truth. These synthetic images were used to provide a reliable and comprehensive framework for quantitative evaluation on this segmentation problem. Results from the submitted methods demonstrate that all the methods are effective in the segmentation of clumps containing at most three cells, with overlap coefficients up to 0.3. This highlights the intrinsic difficulty of this challenge and provides motivation for significant future improvement.

117 citations

Posted Content•
Incorporating prior knowledge in medical image segmentation: a survey.

[...]

Masoud Nosrati, Ghassan Hamarneh
05 Jul 2016-arXiv: Computer Vision and Pattern Recognition
TL;DR: This survey focuses on optimization-based methods that incorporate prior information into their frameworks and reviews and compares these methods in terms of the types of prior employed, the domain of formulation, and the optimization techniques.
Abstract: Medical image segmentation, the task of partitioning an image into meaningful parts, is an important step toward automating medical image analysis and is at the crux of a variety of medical imaging applications, such as computer aided diagnosis, therapy planning and delivery, and computer aided interventions. However, the existence of noise, low contrast and objects' complexity in medical images are critical obstacles that stand in the way of achieving an ideal segmentation system. Incorporating prior knowledge into image segmentation algorithms has proven useful for obtaining more accurate and plausible results. This paper surveys the different types of prior knowledge that have been utilized in different segmentation frameworks. We focus our survey on optimization-based methods that incorporate prior information into their frameworks. We review and compare these methods in terms of the types of prior employed, the domain of formulation (continuous vs. discrete), and the optimization techniques (global vs. local). We also created an interactive online database of existing works and categorized them based on the type of prior knowledge they use. Our website is interactive so that researchers can contribute to keep the database up to date. We conclude the survey by discussing different aspects of designing an energy functional for image segmentation, open problems, and future perspectives.

66 citations

Cites background or methods from "Local optimization based segmentati..."

  • ...Nosrati and Hamarneh(2014) augmented the level set framework with the ability to handle two important and intuitive gometric relationships,containmentand exclusion, along with a distance constraint between boundaries of multi-region objects (Figures16 and 17)....

    [...]

  • ...In energy-based segmentation problems there is a trade-off between fidelity and optimizability (Hamarneh, 2011; McIntosh and Hamarneh, 2012; Ulén et al., 2013; Nosrati and Hamarneh, 2014)....

    [...]

  • ...(Images adopted from (Nosrati and Hamarneh, 2014)) (a) (b) (c) (d) (e) (f) • Skull contains gray matter •Gray matter contains white matter and cerebellum •White matter contains putamen • Putamen and cerebellebum are excluded (g) Figure 17: Brain dPET segmentation....

    [...]

  • ...(Images from (Nosrati and Hamarneh, 2014)) optimality of their solutions, their method can only segment a single object in an image and is limited to handling objects that can be “unfolded” into two coupled surfaces....

    [...]

  • ...Examples of methods that employ this constraint include (Zeng et al., 1998; Goldenberg et al., 2002; Paragios, 2002; Nosrati and Hamarneh, 2014; Nosrati et al., 2013) in the continuous settings, and (Wu et al., 2011; Li et al., 2006; Delong and Boykov, 2009; Ulén et al., 2013) in the discrete…...

    [...]

Journal Article•DOI•
Dual-Channel Active Contour Model for Megakaryocytic Cell Segmentation in Bone Marrow Trephine Histology Images

[...]

Tzu-Hsi Song1, Victor Sanchez1, Hesham EIDaly, Nasir M. Rajpoot1•
University of Warwick1
04 Apr 2017-IEEE Transactions on Biomedical Engineering
TL;DR: A novel framework that can efficiently delineate the nuclei and cytoplasm of these cells in digitized images of bone marrow trephine biopsies is proposed and achieves accurate results for both megakaryocytic nuclear and cy toplasmic delineation.
Abstract: Assessment of morphological features of megakaryocytes (MKs) (special kind of cells) in bone marrow trephine biopsies play an important role in the classification of different subtypes of Philadelphia-chromosome-negative myeloproliferative neoplasms (Ph-negative MPNs). In order to aid hematopathologists in the study of MKs, we propose a novel framework that can efficiently delineate the nuclei and cytoplasm of these cells in digitized images of bone marrow trephine biopsies. The framework first employs a supervised machine learning approach that utilizes color and texture features to delineate megakaryocytic nuclei. It then employs a novel dual-channel active contour model to delineate the boundary of megakaryocytic cytoplasm by using different deconvolved stain channels. Compared to other recent models, the proposed framework achieves accurate results for both megakaryocytic nuclear and cytoplasmic delineation.

46 citations

Cites background or methods from "Local optimization based segmentati..."

  • ...The third set evaluates our complete framework against the multi-region active contour in [41], which allows delineating the nuclear and cytoplasmic boundaries simultaneously....

    [...]

  • ...This map is used along with containment and exclusion energy terms to simultaneously segment multiple regions [41]....

    [...]

  • ...Second, we evaluate the accuracy of our DCAC model in segmenting the corresponding cytoplasmic regions against the CV model, the LBF model, our previously proposed circumscribing active contour (CAC) model [44], and the multi-region active contour model in [41]....

    [...]

  • ...We also compare our entire proposed framework with the multi-region active contour model in [41] using the images and ground truth used in the second experiment....

    [...]

  • ...It is important to mention that in [41], the authors also propose a model that does not favour similar shapes (i....

    [...]

References
More filters
Journal Article•DOI•
Statistical methods for assessing agreement between two methods of clinical measurement.

[...]

J M Bland1, J M Bland2, Douglas G. Altman1, Douglas G. Altman2•
St George's Hospital1, Northwick Park Hospital2
08 Feb 1986-The Lancet
TL;DR: An alternative approach, based on graphical techniques and simple calculations, is described, together with the relation between this analysis and the assessment of repeatability.

43,884 citations

"Local optimization based segmentati..." refers methods in this paper

  • ...12(b), illustrates the Bland–Altman plot [36], which is used to compare two clinical measurements and shows the difference between the two TABLE V LV SEGMENTATION RESULTS (SUNNYBROOK DATASET): AVERAGE DISTANCE ERROR...

    [...]

Journal Article•DOI•
Active contours without edges

[...]

Tony F. Chan1, Luminita A. Vese1•
University of California, Los Angeles1
01 Feb 2001-IEEE Transactions on Image Processing
TL;DR: A new model for active contours to detect objects in a given image, based on techniques of curve evolution, Mumford-Shah (1989) functional for segmentation and level sets is proposed, which can detect objects whose boundaries are not necessarily defined by the gradient.
Abstract: We propose a new model for active contours to detect objects in a given image, based on techniques of curve evolution, Mumford-Shah (1989) functional for segmentation and level sets. Our model can detect objects whose boundaries are not necessarily defined by the gradient. We minimize an energy which can be seen as a particular case of the minimal partition problem. In the level set formulation, the problem becomes a "mean-curvature flow"-like evolving the active contour, which will stop on the desired boundary. However, the stopping term does not depend on the gradient of the image, as in the classical active contour models, but is instead related to a particular segmentation of the image. We give a numerical algorithm using finite differences. Finally, we present various experimental results and in particular some examples for which the classical snakes methods based on the gradient are not applicable. Also, the initial curve can be anywhere in the image, and interior contours are automatically detected.

10,404 citations

"Local optimization based segmentati..." refers methods in this paper

  • ...Assuming piecewise constant regional intensities and adopting the total variation regularization, we employ the standard Chan-Vese formulation [13] to minimize the intra-region variance considering the geometric, e....

    [...]

  • ...Assuming piecewise constant regional intensities and adopting the total variation regularization, we employ the standard Chan-Vese formulation [13] to minimize the intra-region variance considering the geometric, e.g., containment, between regions....

    [...]

  • ...To minimize the functional in (9), we follow the approach of Chan and Vese [13] and derive the Euler-Lagrange equation....

    [...]

Journal Article•DOI•
A Multiphase Level Set Framework for Image Segmentation Using the Mumford and Shah Model

[...]

Luminita A. Vese1, Tony F. Chan1•
University of California, Los Angeles1
01 Dec 2002-International Journal of Computer Vision
TL;DR: A new multiphase level set framework for image segmentation using the Mumford and Shah model, for piecewise constant and piecewise smooth optimal approximations, and validated by numerical results for signal and image denoising and segmentation.
Abstract: We propose a new multiphase level set framework for image segmentation using the Mumford and Shah model, for piecewise constant and piecewise smooth optimal approximations. The proposed method is also a generalization of an active contour model without edges based 2-phase segmentation, developed by the authors earlier in T. Chan and L. Vese (1999. In Scale-Space'99, M. Nilsen et al. (Eds.), LNCS, vol. 1682, pp. 141–151) and T. Chan and L. Vese (2001. IEEE-IP, 10(2):266–277). The multiphase level set formulation is new and of interest on its own: by construction, it automatically avoids the problems of vacuum and overlaps it needs only log n level set functions for n phases in the piecewise constant cases it can represent boundaries with complex topologies, including triple junctionss in the piecewise smooth case, only two level set functions formally suffice to represent any partition, based on The Four-Color Theorem. Finally, we validate the proposed models by numerical results for signal and image denoising and segmentation, implemented using the Osher and Sethian level set method.

2,649 citations

Book•
On the partial difference equations of mathematical physics

[...]

Richard Courant1, Kurt Friedrichs1, Hans Lewy2•
Courant Institute of Mathematical Sciences1, University of California, Berkeley2
04 Sep 2011
TL;DR: In this paper, a discussion of the behavior of the solution as the mesh width tends to zero is presented, and the applicability of the method to more general difference equations and to those with arbitrarily many independent variables is made clear.
Abstract: Problems involving the classical linear partial differential equations of mathematical physics can be reduced to algebraic ones of a very much simpler structure by replacing the differentials by difference quotients on some (say rectilinear) mesh. This paper will undertake an elementary discussion of these algebraic problems, in particular of the behavior of the solution as the mesh width tends to zero. For present purposes we limit ourselves mainly to simple but typical cases, and treat them in such a way that the applicability of the method to more general difference equations and to those with arbitrarily many independent variables is made clear.

2,047 citations

"Local optimization based segmentati..." refers methods in this paper

  • ...To ensure that our level set-based framework is numerically stable, we place an upper bound for the time-step , using the Courant–Friedrichs–Lewy (CFL) condition [17]....

    [...]

Journal Article•DOI•
A Variational Level Set Approach to Multiphase Motion

[...]

Hongkai Zhao1, Tony F. Chan1, Barry Merriman1, Stanley Osher1•
University of California, Los Angeles1
01 Aug 1996-Journal of Computational Physics
TL;DR: In this paper, a coupled level set method for the motion of multiple junctions (of, e.g., solid, liquid, and grain boundaries), which follows the gradient flow for an energy functional consisting of surface tension and bulk energies, is developed.

1,158 citations

"Local optimization based segmentati..." refers methods in this paper

  • ...[71], we enforce an exclusion constraint on two regions by penalizing the area that the two regions share....

    [...]

Related Papers (5)
U-Net: Convolutional Networks for Biomedical Image Segmentation

[...]

05 Oct 2015
Olaf Ronneberger, Philipp Fischer, Thomas Brox
Image segmentation with variational active contours

[...]

01 Jan 2005
Xavier Bresson
Simultaneous multi-object segmentation using local robust statistics and contour interaction

[...]

20 Sep 2010
Yi Gao, Allen Tannenbaum, Ron Kikinis
Midrange Geometric Interactions for Semantic Segmentation

[...]

01 May 2016-International Journal of Computer Vision
Julia Diebold, Claudia Nieuwenhuis, Daniel Cremers
Rule-based image interpretation with models of expected structure

[...]

01 Jul 1990
Stephen Shemlon, Stanley M. Dunn