Automated measurements on cell nuclei in tissue sections. Validation procedures

A methodology was developed for fully automated measurements of nuclear features in Feulgen-stained tissue sections by means of videomicroscopy and image analysis. Segmentation is performed within one minute on 512 X 512 optical density (OD) images covering about 75 nuclei, resulting in a graphic contour overlay. The corresponding image subset is scanned by an object data extraction program, producing the raw figures for statistical interpretation. The segmentation software was evaluated by three tests, involving comparison with manual delineation and assessment of the influence of OD. Two case studies (ACTH-stimulated adrenal cortex and pancreatic carcinoma) illustrate the biologic accuracy and medical significance of the described methodology.     

Automatic segmentation of cell nuclei in Feulgen-stained histological sections of prostate cancer and quantitative evaluation of segmentation results

Digital image analysis of cell nuclei is useful to obtain quantitative information for the diagnosis and prognosis of cancer. However, the lack of a reliable automatic nuclear segmentation is a limiting factor for high-throughput nuclear image analysis. We have developed a method for automatic segmentation of nuclei in Feulgen-stained histological sections of prostate cancer. A local adaptive thresholding with an object perimeter gradient verification step detected the nuclei and was combined with an active contour model that featured an optimized initialization and worked within a restricted region to improve convergence of the segmentation of each nucleus. The method was tested on 30 randomly selected image frames from three cases, comparing the results from the automatic algorithm to a manual delineation of 924 nuclei. The automatic method segmented a few more nuclei compared to the manual method, and about 73% of the manually segmented nuclei were also segmented by the automatic method. For each nucleus segmented both manually and automatically, the accuracy (i.e., agreement with manual delineation) was estimated. The mean segmentation sensitivity/specificity were 95%/96%. The results from the automatic method were not significantly different from the ground truth provided by manual segmentation. This opens the possibility for large-scale nuclear analysis based on automatic segmentation of nuclei in Feulgen-stained histological sections.     

Object Segmentation and Ground Truth in 3D Embryonic Imaging

Many questions in developmental biology depend on measuring the position and movement of individual cells within developing embryos. Yet, tools that provide this data are often challenged by high cell density and their accuracy is difficult to measure. Here, we present a three-step procedure to address this problem. Step one is a novel segmentation algorithm based on image derivatives that, in combination with selective post-processing, reliably and automatically segments cell nuclei from images of densely packed tissue. Step two is a quantitative validation using synthetic images to ascertain the efficiency of the algorithm with respect to signal-to-noise ratio and object density. Finally, we propose an original method to generate reliable and experimentally faithful ground truth datasets: Sparse-dense dual-labeled embryo chimeras are used to unambiguously measure segmentation errors within experimental data. Together, the three steps outlined here establish a robust, iterative procedure to fine-tune image analysis algorithms and microscopy settings associated with embryonic 3D image data sets.     

Automated analysis of muscle fibre images

A technique for studying the distribution and size of different fibre types in muscles is proposed for automated analysis of individual fibres in optical density images from ATPase-stained muscle sections. After delineation, fibres may be classified into different histological types (1, 2A, 2B and 2C) using the measurement of their mean optical density (mOD). The densitometric measurements were obtained from three serial histological slides stained under different conditions. The delineation procedure is performed on one of the images: the resulting mask is fitted to the other images using a linear coordinate transform. Along with densitometric measurements, the lesser diameter of the fibres is computed. Both in processing and in analysis, extensive use was made of mathematical morphology tools. All software was implemented on a VICOM digital image processor, extended with a VISIOMORPH morphoprocessor board.     

Automatic image segmentation of nuclear stained breast tissue sections using color active contour model and an improved watershed method

Automatic image segmentation of immunohistologically stained breast tissue sections helps pathologists to discover the cancer disease earlier. The detection of the real number of cancer nuclei in the image is a very tedious and time consuming task. Segmentation of cancer nuclei, especially touching nuclei, presents many difficulties to separate them by traditional segmentation algorithms. This paper presents a new automatic scheme to perform both classification of breast stained nuclei and segmentation of touching nuclei in order to get the total number of cancer nuclei in each class. Firstly, a modified geometric active contour model is used for multiple contour detection of positive and negative nuclear staining in the microscopic image. Secondly, a touching nuclei method based on watershed algorithm and concave vertex graph is proposed to perform accurate quantification of the different stains. Finally, benign nuclei are identified by their morphological features and they are removed automatically from the segmented image for positive cancer nuclei assessment. The proposed classification and segmentation schemes are tested on two datasets of breast cancer cell images containing different level of malignancy. The experimental results show the superiority of the proposed methods when compared with other existing classification and segmentation methods. On the complete image database, the segmentation accuracy in term of cancer nuclei number is over than 97%, reaching an improvement of 3–4% over earlier methods.     

Automated image-based cytometry with fluorescence-stained specimens

The combination of digitized microscopy, algorithms for object recognition and fluorescent labeling is a promising approach for reliable, quick, automated and cost-effective screening of clinical specimens. We describe two conceptually different algorithms for detecting objects in fluorescence microscopic images. One, which is partially automated, compares a mask that represents a typical object with every position in the image; the other, which is fully automated, calculates threshold intensities to segment the image into regions of objects and background. Applications of the algorithms in conjunction with a prototype image-based cytometer are demonstrated for determining the DNA ploidy distribution of cultured human endometrial cells and determining the DNA ploidy distribution and the fraction of cells expressing the E6 antigen of human papilloma virus serotypes 16 and 18 in a PAP smear. The encouraging results from this study suggest that automated image-based cytometry utilizing fluorescent stains will be a valuable asset for clinical screening.     

A multi-model approach to simultaneous segmentation and classification of heterogeneous populations of cell nuclei in 3D confocal microscope images.

Automated segmentation and morphometry of fluorescently labeled cell nuclei in batches of 3D confocal stacks is essential for quantitative studies. Model-based segmentation algorithms are attractive due to their robustness. Previous methods incorporated a single nuclear model. This is a limitation for tissues containing multiple cell types with different nuclear features. Improved segmentation for such tissues requires algorithms that permit multiple models to be used simultaneously. This requires a tight integration of classification and segmentation algorithms. Two or more nuclear models are constructed semiautomatically from user-provided training examples. Starting with an initial over-segmentation produced by a gradient-weighted watershed algorithm, a hierarchical fragment merging tree rooted at each object is built. Linear discriminant analysis is used to classify each candidate using multiple object models. On the basis of the selected class, a Bayesian score is computed. Fragment merging decisions are made by comparing the score with that of other candidates, and the scores of constituent fragments of each candidate. The overall segmentation accuracy was 93.7% and classification accuracy was 93.5%, respectively, on a diverse collection of images drawn from five different regions of the rat brain. The multi-model method was found to achieve high accuracy on nuclear segmentation and classification by correctly resolving ambiguities in clustered regions containing heterogeneous cell populations.     

Convolutional neural network-based automatic liver delineation on contrast-enhanced and non-contrast-enhanced CT images for radiotherapy planning

AimThis study evaluated a convolutional neural network (CNN) for automatically delineating the liver on contrast-enhanced or non-contrast-enhanced CT, making comparisons with a commercial automated technique (MIM Maestro®).BackgroundIntensity-modulated radiation therapy requires careful labor-intensive planning involving delineation of the target and organs on CT or MR images to ensure delivery of the effective dose to the target while avoiding organs at risk.Materials and MethodsContrast-enhanced planning CT images from 101 pancreatic cancer cases and accompanying mask images showing manually-delineated liver contours were used to train the CNN to segment the liver. The trained CNN then performed liver segmentation on a further 20 contrast-enhanced and 15 non-contrastenhanced CT image sets, producing three-dimensional mask images of the liver.ResultsFor both contrast-enhanced and non-contrast-enhanced images, the mean Dice similarity coefficients between CNN segmentations and ground-truth manual segmentations were significantly higher than those between ground-truth and MIM Maestro software (p < 0.001). Although mean CT values of the liver were higher on contrast-enhanced than on non-contrast-enhanced CT, there were no significant differences in the Hausdorff distances of the CNN segmentations, indicating that the CNN could successfully segment the liver on both image types, despite being trained only on contrast-enhanced images.ConclusionsOur results suggest that a CNN can perform highly accurate automated delineation of the liver on CT images, irrespective of whether the CT images are contrast-enhanced or not.     

Influence of optical density on the automated segmentation of rat kidney lysosomes.

After staining for acid phosphatase, video-images were acquired from 0.5-micron sections of rat kidney. Lysosomes in proximal tubules were automatically segmented, using a VICOM digital image processor and measured for area, number and optical density (OD). The purpose of this study is to objectively evaluate the performance of the automated segmentation algorithm at different staining intensities (a) by measuring area after staining with different incubation times, reduced substrate concentration or by adding an inhibitor and (b) by 'simulating' a decrease in OD (reducing grey-values at each point of a digitized image). The results of the experiments showed that: (1) the algorithm will underestimate the size of lysosomes (a) when the OD in close to the local background and (b) when an area is larger than or close to the area of the lowpass square filter; (2) accuracy of the segmentation can be improved by comparing the results of feature extraction after segmentation of the same image at different relative OD levels; (3) lysosomes with very low OD, compared to background are delineated with a large error or not delineated at all and this cannot be corrected. Incorrectly delineated lysosomes can be identified and excluded from further calculations, or their measured area replaced by an estimate of the true area.     

An evaluation of semiautomatic approaches to contour segmentation applied to fungal hyphae

Semiautomatic image analysis techniques are particularly useful in biological applications, which commonly generate very complex images, and offer considerable flexibility. However, systematic study of such methods is lacking; most research develops fully automatic algorithms. This paper describes a study to evaluate several different semiautomatic or computer-assisted approaches to contour segmentation within the context of segmenting degraded images of fungal hyphae. Four different types of contour segmentation method, with varying degrees and types of user input, are outlined and applied to hyphal images. The methods are evaluated both quantitatively and qualitatively by comparing results obtained by several test subjects segmenting simulated images qualitatively similar to the hyphal images of interest. An active contour model approach, using control points, emerges as the method to be preferred to three more traditional approaches. Feedback from the image provider indicates that any of the methods described have something useful to offer for segmentation of hyphae.     

A Reference-Free Method for Brightness Compensation and Contrast Enhancement of Micrographs of Serial Sections

Three-dimensional (3D) reconstruction of an organ or tissue from a stack of histologic serial sections provides valuable morphological information. The procedure includes section preparation of the organ or tissue, micrographs acquisition, image registration, 3D reconstruction, and visualization. However, the brightness and contrast through the image stack may not be consistent due to imperfections in the staining procedure, which may cause difficulties in micro-structure identification using virtual sections, region segmentation, automatic target tracing, etc. In the present study, a reference-free method, Sequential Histogram Fitting Algorithm (SHFA), is therefore developed for adjusting the severe and irregular variance of brightness and contrast within the image stack. To apply the SHFA, the gray value histograms of individual images are first calculated over the entire image stack and a set of landmark gray values are chosen. Then the histograms are transformed so that there are no abrupt changes in progressing through the stack. Finally, the pixel gray values of the original images are transformed into the desired ones based on the relationship between the original and the transformed histograms. The SHFA is tested on an image stacks from mouse kidney sections stained with toluidine blue, and captured by a slide scanner. As results, the images through the entire stack reveal homogenous brightness and consistent contrast. In addition, subtle color differences in the tissue are well preserved so that the morphological details can be recognized, even in virtual sections. In conclusion, compared with the existing histogram-based methods, the present study provides a practical method suitable for compensating brightness, and improving contrast of images derived from a large number of serial sections of biological organ.     

Segmentation of parotid glands from registered CT and MR images

PurposeTo develop an automatic multimodal method for segmentation of parotid glands (PGs) from pre-registered computed tomography (CT) and magnetic resonance (MR) images and compare its results to the results of an existing state-of-the-art algorithm that segments PGs from CT images only.MethodsMagnetic resonance images of head and neck were registered to the accompanying CT images using two different state-of-the-art registration procedures. The reference domains of registered image pairs were divided on the complementary PG regions and backgrounds according to the manual delineation of PGs on CT images, provided by a physician. Patches of intensity values from both image modalities, centered around randomly sampled voxels from the reference domain, served as positive or negative samples in the training of the convolutional neural network (CNN) classifier. The trained CNN accepted a previously unseen (registered) image pair and classified its voxels according to the resemblance of its patches to the patches used for training. The final segmentation was refined using a graph-cut algorithm, followed by the dilate-erode operations.ResultsUsing the same image dataset, segmentation of PGs was performed using the proposed multimodal algorithm and an existing monomodal algorithm, which segments PGs from CT images only. The mean value of the achieved Dice overlapping coefficient for the proposed algorithm was 78.8%, while the corresponding mean value for the monomodal algorithm was 76.5%.ConclusionsAutomatic PG segmentation on the planning CT image can be augmented with the MR image modality, leading to an improved RT planning of head and neck cancer.     

Segmentation of the Clustered Cells with Optimized Boundary Detection in Negative Phase Contrast Images

Cell image segmentation plays a central role in numerous biology studies and clinical applications. As a result, the development of cell image segmentation algorithms with high robustness and accuracy is attracting more and more attention. In this study, an automated cell image segmentation algorithm is developed to get improved cell image segmentation with respect to cell boundary detection and segmentation of the clustered cells for all cells in the field of view in negative phase contrast images. A new method which combines the thresholding method and edge based active contour method was proposed to optimize cell boundary detection. In order to segment clustered cells, the geographic peaks of cell light intensity were utilized to detect numbers and locations of the clustered cells. In this paper, the working principles of the algorithms are described. The influence of parameters in cell boundary detection and the selection of the threshold value on the final segmentation results are investigated. At last, the proposed algorithm is applied to the negative phase contrast images from different experiments. The performance of the proposed method is evaluated. Results show that the proposed method can achieve optimized cell boundary detection and highly accurate segmentation for clustered cells.     

Evaluating the performance of microarray segmentation algorithms

MOTIVATION: Although numerous algorithms have been developed for microarray segmentation, extensive comparisons between the algorithms have acquired far less attention. In this study, we evaluate the performance of nine microarray segmentation algorithms. Using both simulated and real microarray experiments, we overcome the challenges in performance evaluation, arising from the lack of ground-truth information. The usage of simulated experiments allows us to analyze the segmentation accuracy on a single pixel level as is commonly done in traditional image processing studies. With real experiments, we indirectly measure the segmentation performance, identify significant differences between the algorithms, and study the characteristics of the resulting gene expression data. RESULTS: Overall, our results show clear differences between the algorithms. The results demonstrate how the segmentation performance depends on the image quality, which algorithms operate on significantly different performance levels, and how the selection of a segmentation algorithm affects the identification of differentially expressed genes. AVAILABILITY: Supplementary results and the microarray images used in this study are available at the companion web site http://www.cs.tut.fi/sgn/csb/spotseg/     

Square-cut: a segmentation algorithm on the basis of a rectangle shape

We present a rectangle-based segmentation algorithm that sets up a graph and performs a graph cut to separate an object from the background. However, graph-based algorithms distribute the graph's nodes uniformly and equidistantly on the image. Then, a smoothness term is added to force the cut to prefer a particular shape. This strategy does not allow the cut to prefer a certain structure, especially when areas of the object are indistinguishable from the background. We solve this problem by referring to a rectangle shape of the object when sampling the graph nodes, i.e., the nodes are distributed non-uniformly and non-equidistantly on the image. This strategy can be useful, when areas of the object are indistinguishable from the background. For evaluation, we focus on vertebrae images from Magnetic Resonance Imaging (MRI) datasets to support the time consuming manual slice-by-slice segmentation performed by physicians. The ground truth of the vertebrae boundaries were manually extracted by two clinical experts (neurological surgeons) with several years of experience in spine surgery and afterwards compared with the automatic segmentation results of the proposed scheme yielding an average Dice Similarity Coefficient (DSC) of 90.97±2.2%.     

Segmentation des images TEP au 18F-FDG. Principe et revue de la littérature

Several segmentation methods of lesion uptake in ¹⁸F-FDG PET imaging have been proposed in the literature. Their principles are presented along with their clinical results. The main approach proposed in the literature is the thresholding method. The most commonly used is a constant threshold around 40% of the maximum uptake within the lesion. This simple approach is not valid for small (< 4 or 5 mL), poorly contrasted positive tissue (SUV < 2) or lesion in movement. To limit these problems, more complex thresholding algorithms have been proposed to define the optimal threshold value to be applied to segment the lesion. The principle is to adapt the threshold following a fitting model according to one or two characteristic image parameters. Those algorithms based on iterative approaches to find the optimal threshold value are preferred as they take into account patient data. The main drawback is the need of a calibration step depending on the PET device, the acquisition conditions and the algorithm used for image reconstruction. To avoid this problem, some more sophisticated segmentation methods have been proposed in the literature: derivative methods, watershed and pattern recognition algorithms. The delineation of positive tissue on FDG-PET images is a complex problem, always under investigation.     

A microcomputer based system for three-dimensional reconstructions from tomographic or histologic sections

The reconstructions of three-dimensional (3-D) objects from serial two-dimensional (2-D) images can contribute to the understanding of many biologic structures, from organelles to organs and tissues. The 3-D reconstruction of sections can be divided into several major tasks: image acquisition, alignment of slices, internal object definition, object reconstruction and rotation of the completed image. A fast, versatile, interactive system was devised for the reconstruction of 3-D objects from serial 2-D images using a low-cost microcomputer, original programs and commercial software. The system allows reconstruction from any serial images, e.g., electron micrographs, histologic sections or computed tomograms. A photographic image or a microscopic field is acquired into the computer memory using a video digitizer. Slices are superimposed and aligned to each other using an operator-interactive program. A contour-(edge-) finding algorithm isolates an object of interest from the background image by "subtraction" of the image from an overlaid, slightly shifted identical image. Contours for each slice are input to a reconstruction procedure, which calculates the x, y and z coordinates of every point in a slice and the thickness and number of slices. It then calculates the illumination for every point using a given point source of light and an intensity-fading coefficient. Finally, the points are represented by cubes to provide dimension and reflective surfaces. A cube of appropriate shade and color represents in 2-D the equivalent of a 3-D object; this results in a very effective 3-D image. The reconstruction is rotated by recalculating the positions of every point defining the object and rebuilding the image.(ABSTRACT TRUNCATED AT 250 WORDS)     

Low-level aspects of segmentation and recognition.

This paper discusses two problems related to three-dimensional object recognition. The first is segmentation and the selection of a candidate object in the image, the second is the recognition of a three-dimensional object from different viewing positions. Regarding segmentation, it is shown how globally salient structures can be extracted from a contour image based on geometrical attributes, including smoothness and contour length. This computation is performed by a parallel network of locally connected neuron-like elements. With respect to the effect of viewing, it is shown how the problem can be overcome by using the linear combinations of a small number of two-dimensional object views. In both problems the emphasis is on methods that are relatively low level in nature. Segmentation is performed using a bottom-up process, driven by the geometry of image contours. Recognition is performed without using explicit three-dimensional models, but by the direct manipulation of two-dimensional images.