Abstract: Cell shape is an important characteristic of the physiological state of a cell and is used as a primary read-out of cell behaviour in various assays. Automated accurate segmentation of cells in microscopy images is hence of large practical importance in cell biology. We report a simple algorithm for automated cell segmentation in high-magnification phase-contrast images, which takes advantage of the characteristic directionality of the local image intensity gradient at cellular boundaries due to the 'halo-effect'. We employ a two-step algorithm in which a gradient vector flow (GVF) field is first used to direct active contours to an approximate cell boundary. A directional GVF (DGVF) field is then calculated by considering only edges for which the image intensity gradient is directed outwards with respect to the approximate cell contour. Subsequently, the DGVF field is used to refine the cell contour, by directing active contours to edges with the desired gradient directionality. This method allows us to accurately segment cells in an image series, as well as follow the dynamics of cell shape over time in an automated fashion.

Abstract: In this paper a left ventricle segmentation approach in short-axis MRI is proposed. It is based on an active contour method and gradient vector flow field forces. Firstly, algorithm delineates endocardium using active contour method approach assisted by gradient vector flow field forces. After that, the epicardium is outlined by proposed divergence rays method and corrected by Fourier descriptors to smoothen an epicardium curve. An algorithm has been tested on eight healthy patients and compared to a manual delineation of endo- and epicardium boundaries. Validity of an algorithm is checked by linear regression analysis, correlation coefficients, and RSME errors. Sample Pearson product-moment correlation coefficients between automatic and manual delineation are rENDO=0.95 and rEPI=0.86. The coefficients of determination and RMSEs are $R^2_{ENDO}=0.9$, $R^2_{EPI}=0.74$ and $RMSE_{ENDO}=5.303 \ ml$, $RMSE_{EPI}=21.973 \ ml$, respectively. These experiments confirm accuracy and robustness of the proposed approach.

Abstract: Manual segmentation of brain tumors by medical practitioners is a time consuming task and has inability to assist in accurate diagnosis. Several automatic methods have been developed to overcome these issues. But Automatic MRI (Magnetic Resonance Imaging) brain tumor segmentation is a complicated task due to the variance and intricacy of tumors; to over by this problem we have developed a new method for automatic classification of brain tumor. In the proposed method the MRI Brain image classification of tumors is done based on Fluid vector flow and support vector machine classifier. In this method Fluid Vector Flow is utilized for segmentation of two dimensional brain tumor MR images to extract the tumor and that tumor can be projected into the three dimensional plane to analyze the depth of the tumor. Finally, Support vector machine classifier is utilized to perform two functions. The first is to differentiate between normal and abnormal. The second function is to classify the type of abnormality in benign or malignant tumor.

Abstract: Conventional pulse wave Doppler techniques can only provide one-dimensional blood velocity components parallel to the direction of the beam and conventional focusing provides limited frame rates of about 30-40 frames per second. As a solution to these well known limitations we perform a two-dimensional vector mapping using compounded coplanar oriented plane waves, analogous to vector-Doppler. Our method was tested using Field II simulations of both stationary parabolic pipe flow and computational fluid dynamics determined flow through a patient specific carotid artery. Our results show the ability for this method to provide more discernible representation of the flow dynamics compared with conventional color-Doppler imaging, while maintaining a frame rate of roughly 500 frames per second. Quantitative comparison with known velocity fields provides robust validation and demonstrates error comparable to that found in literature using conventional Doppler measurements. Moreover, this method provides a promising means to quantify quick transitory events and complex flow structures unattainable with clinical color-Doppler.

Abstract: Vector velocity blood flow imaging gives speed and direction of blood flow at each pixel. An imaging algorithm proposed earlier [2] requires multiple angles of planewave (PW) transmissions to construct a robustly invertible model for vector velocity estimates. Here we demonstrate a vector velocity estimation approach that requires only a single planewave transmission angle. The proposed algorithm uses PW transmission and reconstruction to generate a blood motion image sequence in the B-mode flow (B-Flow) modality, at frame rates in the Doppler PRF regime. Pixel ensembles in the image sequence at point p = [x, z] and pulse t are comprised of IQ magnitude values, computed from the IQ data at each pixel p after wall filtering the ensemble. The sequence of values thus captures motion at a framerate equal to the PRF, revealing fine-scale flow dynamics as a moving texture in the blood reflectivity. Using the chain rule, spatial and temporal derivatives resulting from the space-time gradient of the image sequence couple to the texture flow velocity vector field [vx(x, z, t), vz(x, z, t)] at each pixel p and PRI t. The resulting Gauss-Markov models are solved by least squares to give the vector velocity estimates, which are formulated in the model to be constant over the estimation window. We also evaluate variants that include in the observation, lag-product samples (autocorrelation summands) at non-zero lags, as well as instantaneous Doppler-derived axial velocity estimates. Compared to the multi-angle planewave algorithm presented in [2], this approach allows for a longer time interval for wall filtering, as the frame is not partitioned into separate segments for different planewave angles. This permits wall filters with steeper transition bands, and allows flexibility in balancing framerate and sensitivity, suggesting application to vector flow imaging of deep tissue where efficiently achieving planewave angle diversity at the target becomes difficult. Using a Philips L7-4 transducer and a Verasonics (TM) acquisition system, we evaluate single-angle PWT vector velocity imaging on a Doppler string phantom, and demonstrate it successfully on a carotid artery.

Abstract: Active contour model has a good performance in consecutive boundary extraction for medical images The gradient vector flow (GVF) field is one of the most popular external forces that can increase the capture range and converge to concavities, although it is sensitive to image noise and easy to leak in weak edge Here we propose a novel sigmoid gradient vector flow (SGVF) force model for improving contour performance This novel external force field is insensitive to noises and may prevent the weak edge leakage To further illustrate the advantages associated with the proposed GVF field formulation, synthetic images and real images are conducted when the proposed method is applied in ultrasound image and magnetic resonance image for suppressing noise and extracting the weak edges Experimental results demonstrate that the proposed method leads to more accurate segmentation

Abstract: We propose a novel external force for active contours, which we call neighborhood-extending and noise-smoothing gradient vector flow (NNGVF). The proposed NNGVF snake expresses the gradient vector flow (GVF) as a convolution with a neighborhood-extending Laplacian operator augmented by a noise-smoothing mask. We find that the NNGVF snake provides better segmentation than the GVF snake in terms of noise resistance, weak edge preservation, and an enlarged capture range. The NNGVF snake accomplishes this with a reduced computational cost while maintaining other desirable properties of the GVF snake, such as initialization insensitivity and good convergences at concavities. We demonstrate the advantages of NNGVF on synthetic and real images.

Abstract: In this paper, we propose a novel method for text detection in natural scenes. Gradient Vector Flow is first used to extract both intra-character and inter-character symmetries. In the second step, we group horizontally aligned symmetry components into text lines based on several constraints on sizes, positions and colors. Finally, to remove false positives, we employ a learning-based approach which makes use of Histogram of Oriented Gradients feature. The main advantage of the proposed method lies in the use of both the text features and the gap (i.e., inter-character) features. Existing techniques typically extract only the former and ignore the latter. Experiments on the benchmark ICDAR 2003 dataset show the good detection performance of our method on natural scene text.

Abstract: In this paper we present a new method to generate the external force field for the active contour. The new approach, multi-scale vector field convolution (MVFC), is based on the vector field convolution (VFC) method, and extends VFC to a multi-scale process. Via automatic scale selection, the proposed method constructs the external force field used by the active contour by varying the convolution kernel width according to local image structure. The result is an adaptive force generation method that accommodates detailed, high curvature features with a large capture range, and demonstrates improved performance in terms of local convergence and multi-target segmentation over traditional VFC. Meanwhile, the new method preserves the low computation complexity as supported by VFC and exemplifies as a more efficient force generation method than gradient vector flow (GVF). Synthetic and real experiments demonstrate the efficacy of MVFC in terms of detail preservation, capture range, noise resilience and computational cost.

Abstract: To obtain hemodynamic-relevant parameters in case of cardiovascular diseases the velocity-encoded magnetic resonance imaging (PC-MRI) is used for the non-invasive measurement of the blood flow in terms of 3D velocity fields. During the segmentation of the vessel lumen in those datasets conventional segmentation methods often fail due to reduced image quality. In this paper we present a method for the centerline extraction of great vessels in PC-MR images using additional features extracted from vector flow information. The proposed algorithm can be divided in the following steps: the propagation along the vessel course by using streamlines and the largest eigenvector, the radial search for the vessel boundary, the determination of the center position in the cross-sectional plane of the vessel and the adjustment of the propagation step size subject to the vessel curvature. This is done by using a combination of morphology and flow information: the Sobel filtered and the threshold filtered image as morphologic features as well as the coherence values of the flow vectors and the behaviour of the blood flow streamlines within the vessel and around the borders as flow features. The developed algorithm was evaluated on clinical PC-MRI datasets with encouraging results. The centerline points of the entire aorta as well as corresponding border points were successfully extracted for 16 out of 17 examined datasets. For the detection of the vessel boundary the features extracted from flow information showed to yield more reliable results than morphology features.

Abstract: An ultrasound imaging console includes receive circuitry that receives a set of echoes produced in response to an ultrasound signal traversing blood flowing in a portion of a vessel in a field of view, a beamformer that beamforms the echoes, a velocity processor that determines flow direction and magnitude of the flowing blood based on the beamformed echoes, and a rendering engine that displays the determined flow direction and magnitude.

Abstract: Medical image segmentation finds application in computer-aided diagnosis, computer-guided surgery, measuring tissue volumes, locating tumors, and pathologies. One approach to segmentation is to use active contours or snakes. Active contours start from an initialization (often manually specified) and are guided by image-dependent forces to the object boundary. Snakes may also be guided by gradient vector fields associated with an image. The first main result in this direction is that of Xu and Prince, who proposed the notion of gradient vector flow (GVF), which is computed iteratively. We propose a new formalism to compute the vector flow based on the notion of bilateral filtering of the gradient field associated with the edge map — we refer to it as the bilateral vector flow (BVF). The range kernel definition that we employ is different from the one employed in the standard Gaussian bilateral filter. The advantage of the BVF formalism is that smooth gradient vector flow fields with enhanced edge information can be computed noniteratively. The quality of image segmentation turned out to be on par with that obtained using the GVF and in some cases better than the GVF.

Abstract: An experimental validation system for vector-flow-mapping (VFM) has been developed to verify the VFM accuracy. A VFM measured by ultrasound is a mapping of 2D flow vector fields in the human heart. The validation system consists of a left ventricle (LV) phantom, a stereo particle image velocimetry (stereo-PIV), and an ultrasound scanner. The LV phantom is pulsated by controlling the external pressure. The optically transparent LV phantom enables the PIV to obtain velocity fields. To evaluate the degradation of VFM accuracy in a 3D flow, the VFM results are compared with those obtained from the stereo-PIV, which provides accurate 3D velocity components in a plane. Preliminary results suggest that the VFM and PIV vector fields are in good agreement. However, 3D flow, which resulted from the VFM errors, are also observed in the phantom. A novel method for estimating uncertainty in individual VFM measurements was proposed by considering the velocity discrepancy between VFM and tissue tracking at the LV wall. The VFM error estimated with the proposed method was validated by comparing it with that measured using the PIV. The estimated errors agreed with the measured error from the PIV. The mean difference was about 8% in the experiments.

Abstract: The incorrect shape and location of initial contour in the level set image segmentation method could lead to unreliable results but the proper initial contour is hard to design in some images This paper proposes a method that is more independent using four vector flows instead of using edge detector to insert new contour The geodesic active contour (GAC) is used to compared with the proposed method on several types of gray-scale images The experimental results show that the proposed method is more independent from shape and location of initial contour and can cover more details comparing with GAC

Abstract: A novel external force for active contours, called as adaptive vector flow (AVF), is proposed in this paper. Based on analyzing the diffusion mechanism of gradient vector flow (GVF), it is found that GVF is difficult to preserve weak edges and enter long and thin concavities. In AVF, we replace the isotropic smoothness term of GVF by an adaptive anisotropic one and adjust the diffusion speed in tangent and normal directions by the local features of the images. Experimental results on synthetic and real images show that, compared with the GVF snake, the AVF snake has better performance and properties.

Abstract: The use of parallel receive beamforming in conjunction with plane wave imaging resulted in a robust vector Doppler modality, applicable for a wide range of patients. A sufficient B-mode quality was maintained by adopting a packet based acquisition scheme with a separate setup for each modality. Retrospective Pulsed Wave (PW) Doppler based on the same recording was feasible, and when combined with the available vector flow information, calibrated velocity spectra could be generated from arbitrary points in the image. All together, the proposed approach may provide more efficient clinical tools for conventional vascular imaging, as well as quantitative information for research into new markers for cardiovascular diagnosis.

Abstract: This paper proposes a fully automatic active-contour-based segmentation method, for the detection of the carotid artery wall in longitudinal B-mode images. A multiscale edge detection methodology is used for the definition of an initial snake contour, followed by a gradient vector flow (GVF) snake. The multiscale edge detection method is based on finding the local maxima of the wavelet transform of the image, which is very close to the real contour. The GVF snake is based on the calculation of the image edge map and the calculation of the GVF field which guides the deformation for the estimation of the real arterial wall boundaries. In twenty cases of healthy carotid arteries the sensitivity, specificity and accuracy were higher than 0.97, 0.99 and 0.98 respectively, for both diastolic and systolic cases. In conclusion, the proposed methodology provides a new accurate procedure to detect the arterial wall in ultrasound images of the carotid artery.

Abstract: Gradient vector flow(GVF) is an effective external force for active contours,but its isotropic nature handicaps its performance.The recently proposed gradient vector flow in the normal direction(NGVF) is anisotropic since it only keeps the diffusion along the normal direction of the isophotes;however,it has difficulties forcing a snake into long,thin boundary indentations.In this paper,a novel external force for active contours called normally generalized gradient vector flow(NGGVF) is proposed,which generalizes the NGVF formulation to include two spatially varying weighting functions.Consequently,the proposed NGGVF snake is anisotropic and would improve active contour convergence into long,thin boundary indentations while maintaining other desirable properties of the NGVF snake,such as enlarged capture range,initialization insensitivity and good convergence at concavities.The advantages on synthetic and real images are demonstrated.

Abstract: This PhD thesis has investigated the use of a new ultrasound technique that to measure the movement of blood. The technique was developed at the Center for Fast Ultrasound Imaging at the Technical University of Denmark and has previously only been available with experimental ultrasound scanners. Now, the method has been implemented into a commercial ultrasound scanner made for hospital use. In real-time, the technique measures movements in all directions as 2D vector fields, including movements perpendicular to the ultrasound beam. This is not available with conventional ultrasound scanners today. The thesis consists of three studies that uses vector flow ultrasound measurements on healthy volunteers. In study I the common carotid artery of 16 healthy volunteers were scanned simultaneously with the vector technique and the conventional, spectral estimation method. The study compared the clinical parameters: peak systole velocity, end diastole velocity, resistive index, and the flow direction. The results showed significant difference on the velocities and the resistive index. However, no significant difference on the manually defined flow angle and the calculated mean flow angle by the vector technique. With the conventional technique, the manual setting of the angle is operator dependent. With the calculated vector angle, this operator is relieved from the angle setting and the measurement is angle corrected by the identical method every time. With study II the carotid bifurcation including the carotid bulb and the common carotid artery were scanned on 8 healthy volunteers. The flow patterns of the two structures were outlined and presented to each of 5 experienced radiologists. The complexity of the identical areas were calculated by the vector concentration and compared to the visual evaluations. No significant difference was found between the two methods which were equally good at discriminating the laminar flow of the common carotid artery from the complex flow in the carotid bulb. Thus, a new method was presented to quantify complex flow patterns with vector flow. The final study III presented the rotational flow patterns in the cross-sectional plane of three arteries: The common carotid artery, the abdominal aorta, and the common iliac artery. Five healthy volunteers were included in the study and nine datasets visualized the flow patterns during the diastole. The rotational frequency was calculated and the results indicate a constant direction of the rotation for each artery. Extended measurements on the abdominal aorta showed a two-directional rotation during the cardiac cycle. An observation that corresponds to previous MR and Doppler studies. With the three studies, this thesis presents new methods that quantifies in vivo vector flow obtained in real-time with a new implementation.

Abstract: Concerning that the edges of the endocardium and epicardium of the left ventricle in the cardiac Magnetic Resonance Imaging(MRI) images have different directions,a new directional active contour model in curve evolution framework was proposed for segmentation of endocardium and epicardium of the left ventricle.The curve evolution equation included a hybrid geometric flow with edge and region gray characteristics that were obtained from the image itself.The edge-based term in the geometric flow borrowed from extended Dynamic Directional Gradient Vector Flow(DDGVF) with fast marching method was utilized to guide the curve evolution towards the object boundaries with different direction.The region-based term borrowed from Chan-Vese(CV) model was utilized to prevent the curve from leakage under the influence of other edge components.The final curve evolution equation was dealt with level set method.The experimental results for gray and cardiac MRI images show that the proposed method can get better segmentation effects.It has certain application value for realizing myocardium auto-segmentation,evaluation and analysis of heart function based on cardiac MRI images.

Abstract: A novel external force called extended neighborhood generalized gradient vector flow ( ENGGVF ) for active contours is proposed in this paper. The proposed ENGGVF snake incorporates weighting functions and takes the gradient vector flow (GVF) just as a convolution operation with extended neighborhood Laplacian operator mask. Consequently, the ENGGVF snake would provide much better segmentation than GVF snake in terms of deep and narrow concavity convergence, reduced computational cost and enlarged capture range while maintaining other desirable properties of GVF snake such as initialization insensitivity and U-shape concavity convergence. We demonstrate the advantages on synthetic and real images.

Abstract: Gradient vector flow(GVF) is an effective external force for active contours,but its isotropic nature handicaps its performance.The recently proposed gradient vector flow in the normal direction(NGVF) is anisotropic since it only keeps the diffusion along the normal direction of the isophotes;however,it has difficulties forcing a snake into long,thin boundary indentations.In this paper,a novel external force for active contours called normally generalized gradient vector flow(NGGVF) is proposed,which generalizes the NGVF formulation to include two spatially varying weighting functions.Consequently,the proposed NGGVF snake is anisotropic and would improve active contour convergence into long,thin boundary indentations while maintaining other desirable properties of the NGVF snake,such as enlarged capture range,initialization insensitivity and good convergence at concavities.The advantages on synthetic and real images are demonstrated.

Abstract: Gradient Vector Flow has become a popular method to recover medial information in medical imaging, in particular for vessels centerline extraction. This renewed interest has been motivated by its ability to process gray-scale images without prior segmentation. However, another interesting property lies in the diffusion process used to solve the underlying variational problem. We propose a method to recover scale information in the context of vascular structures extraction, relying on analytical properties of the Gradient Vector Flow only, with no multiscale analysis. Through simple one-dimensional considerations, we demonstrate the ability of our approach to estimate the radii of the vessels with an error of 10% only in the presence of noise and less than 3% without noise. Our approach is evaluated on convolved bar-like templates and is illustrated on 2D X-ray angiographic images.

Abstract: The paper presents two new modifications of the Generalized Gradient Vector Flow Snakes based on the directional analysis of the corresponding vector field.

Abstract: A novel external force field for active contours, called gradient vector flow based on anisotropic diffusion (ADGVF), is proposed in this paper. The generation of ADGVF contains an anisotropic diffusion process that the diffusion in the tangent and normal directions to the isophote lines has different diffusion speeds which are locally adjusted according the local structures of the image. The proposed method can address the problem associated with poor convergence of gradient vector flow in the normal direction (NGVF) to the long, thin boundary indentations and the openings of the boundaries. It can improve active contour convergence to these positions. In its numerical implementation, an efficient numerical schema is used to ensure sufficient numerical accuracy. Experimental results demonstrate that ADGVF has better performance in terms of accuracy, efficiency and robustness that that of NGVF.

Abstract: Active contours are one of the most successful image segmentation methods in image processing and computer vision field. Their main limitations are high noise sensitivity and poor capture range from the target object. One of the most promising approaches for solving these limitations is the gradient vector flow (GVF). However, GVF still has improving space in converging to concavities and noise robustness. Here we propose a novel GVF external force based on modified normal flow for improving contour performance. This novel external force field is insensitive to noises and may converge to concavities. We compared the proposed method with other methods by synthetic images and real medical images. Experimental results illustrated that the proposed method had achieved more accurate segmentation for noise robustness and concavity convergence.

Abstract: In this paper, we consider a new length preserving curve flow for closed convex curves in the plane. We show that the flow exists globally, the area of the region bounded by the evolving curve is increasing, and the evolving curve converges to the circle in C ∞ topology as t → ∞.