Contour Generator Research Articles

CylinGCN: Cylindrical structures segmentation in 3D biomedical optical imaging by a contour-based graph convolutional network

Cylindrical organs, e.g., blood vessels, airways, and intestines, are ubiquitous structures in biomedical optical imaging analysis. Image segmentation of these structures serves as a vital step in tissue physiology analysis. Traditional model-driven segmentation methods seek to fit the structure by constructing a corresponding topological geometry based on domain knowledge. Classification-based deep learning methods neglect the geometric features of the cylindrical structure and therefore cannot ensure the continuity of the segmentation surface. In this paper, by treating the cylindrical structures as a 3D graph, we introduce a novel contour-based graph neural network for 3D cylindrical structure segmentation in biomedical optical imaging. Our proposed method, which we named CylinGCN, adopts a novel learnable framework that extracts semantic features and complex topological relationships in the 3D volumetric data to achieve continuous and effective 3D segmentation. Our CylinGCN consists of a multiscale 3D semantic feature extractor for extracting inter-frame multiscale semantic features, and a residual graph convolutional network (GCN) contour generator that combines the semantic features and cylindrical topological priors to generate segmentation contours. We tested the CylinGCN framework on two types of optical tomographic imaging data, small animal whole body photoacoustic tomography (PAT) and endoscopic airway optical coherence tomography (OCT), and the results show that CylinGCN achieves state-of-the-art performance. Code will be released at https://github.com/lzc-smu/CylinGCN.git.

Detection of artificial fragments embedded in remote sensing images by adversarial neural networks

We investigate algorithms for detecting artificial fragments of remote sensing images generated by adversarial neural networks. We consider a detector of artificial images based on the detection of a spectral artifact of generative-adversarial neural networks that is caused by a layer for enhancing the resolution. We use the detecting algorithm to detect artificial fragments embedded in natural remote sensing images using an adversarial neural network that includes a contour generator. We use remote sensing images of various types and resolutions, whereas the substituted areas, some being not simply connected, have different sizes and shapes. We experimentally prove that the investigated spectral neural network detector has high efficiency in detecting artificial fragments of remote sensing images.

Developable surfaces along frontal curves on embedded surfaces

We consider two types of developable surfaces along a frontal curve on an embedded surface in the Euclidean 3-space. One is called the osculating developable surface, and the other is called the normal developable surface. The frontal curve may have singular points. We give new invariants of the frontal curve which characterize singularities of the developable surfaces. Moreover, a frontal curve is a contour generator with respect to an orthogonal projection or a central projection if and only if one of these invariants is constantly equal to zero.

Flat Approximations of Surfaces Along Curves

Abstract We consider a developable surface tangent to a surface along a curve on the surface. We call it an osculating developable surface along the curve on the surface. We investigate the uniqueness and the singularities of such developable surfaces. We discover two new invariants of curves on a surface which characterize these singularities. As a by-product, we show that a curve is a contour generator with respect to an orthogonal projection or a central projection if and only if one of these invariants constantly equal to zero.

A Novel Method in Extracranial Removal of Brain MR Images

The removal of extracranial elements from brain M R images provides doctors a reliable method to analyze and diagnose dynamic data in E-Health. In traditional segmentation, experts are required to identify every element of the whole image, subjectivity and large amount of works make the mission uncertain and intolerant. GVF Snake model leads to efficiently dealing with the segmentation by giving an initial contour around the final object, but manual participation is still inevitable. This paper proposes an automatic morphology -based algorithm to generate the initial contour for active contour model to implement the removal accurately. To achieve the removal, initial contours will be produced by the original contour generator, which are utilized to approach more precise contours implemented by improved GVF Snake model. After a self-correction among the elements, segmentation missions are done. Experimental result shows that with simple steps and little time, the proposed algorithm can complete the segmentation task successfully, and is of good robustness as well as high accuracy.

Active Contour without Edges vs GVF Active Contour for Accurate Pupil Segmentation

Iris localization is a critical step for an iris recognition system because it directly affects the recognition rates. Consequently, in order to have reasonably accurate measures, we should estimate as many iris boundaries as possible which are defined by papillary and ciliary regions. Due to the contraction which is an intrinsic propriety of the pupil and the variations in the shooting angle, the pupil will not be a regular circle. So an active contour is suitable to accurately locate the iris boundaries. In this paper we focused on iris/pupil boundary and we proposed a new algorithm based on an active contour without edges applied in gray level image. First, we develop a new method to locate and fill the corneal reflection which is used not only to remove the highlight points that appear inside the pupil but also as an initial contour generator for the snake. Second, we propose to use the active contour without edges for precise pupil segmentation. This kind of snake can detect objects whose boundaries are not necessarily defined by gradient. Our algorithm seems to be robust to occlusion, specular reflection, variation in illumination and improves its efficiency in precision and time computation compared with AIPF and Gvf active contour. Another advantage is that the initial curve can be anywhere in the image and the contour will be automatically detected. The proposed algorithm is 2.36 faster than GVF snake-based method for accurate pupil contour detection and integrodifferential method with accuracy up to 99.62% using CASIA iris database V3.0 and up to 100% with CASIA iris database V1.0.

3D surface point and wireframe reconstruction from multiview photographic images

This paper describes a new method for reconstructing 3D surface points and a wireframe on the surface of a freeform object using a small number, e.g. 10, of 2D photographic images. The images are taken at different viewing directions by a perspective camera with full prior knowledge of the camera configurations. The reconstructed surface points are frontier points and the wireframe is a network of contour generators. Both of them are reconstructed by pairing apparent contours in the 2D images. Unlike previous works, we empirically demonstrate that if the viewing directions are uniformly distributed around the object’s viewing sphere, then the reconstructed 3D points automatically cluster closely on a highly curved part of the surface and are widely spread on smooth or flat parts. The advantage of this property is that the reconstructed points along a surface or a contour generator are not under-sampled or under-represented because surfaces or contours should be sampled or represented with more densely points where their curvatures are high. The more complex the contour’s shape, the greater is the number of points required, but the greater the number of points is automatically generated by the proposed method. Given that the viewing directions are uniformly distributed, the number and distribution of the reconstructed points depend on the shape or the curvature of the surface regardless of the size of the surface or the size of the object. The unique pattern of the reconstructed points and contours may be used in 3D object recognition and measurement without computationally intensive full surface reconstruction. The results are obtained from both computer-generated and real objects.

Computing contour generators of evolving implicit surfaces

The contour generator is an important visibility feature of a smooth object seen under parallel projection. It is the curve on the surface which seperates front-facing from back-facing regions. The apparent contour is the projection of the contour generator onto a plane perpendicular to the view direction. Both curves play an important role in computer graphics.Our goal is to obtain fast and robust algorithms that compute the contour generator with a guarantee of topological correctness. To this end, we first study the singularities of the contour generator and apparent contour for both generic views and generic time-dependent projections, for example, when the surface is rotated or deformed. The singularities indicate when components of the contour generator merge or split as time evolves.We present an algorithm to compute an initial contour generator by using a dynamic step size. An interval test guarantees the topological correctness. This initial contour generator can thus be maintained under a time-dependent projection by examining its singularities.

Visual Motion of Curves and Surfaces

For smooth curved surfaces the dominant image feature is the apparent contour, or outline. This is the projection of the contour generator, the locus of points on the surface which separate visible and occluded parts. The contour generator is dependent of the local surface geometry and the viewpoint. Each viewpoint will generate a different contour generator. This paper addresses the problem of recovering the three–dimensional shape and motion of curves and surfaces from image sequences of apparent contours. For known viewer motion the visible surfaces can then be reconstructed by exploiting a spatio–temporal parametrization of the apparent contours and contour generators under viewer motion. A natural parametrization exploits the contour generators and the epipolar geometry between successive viewpoints. The epipolar parametrization leads to simplified expressions for the recovery of depth and surface curvatures from image velocities and accelerations and known viewer motion. The parametrization is, however, degenerate when the apparent contour is singular since the ray is tangent to the contour generator and at frontier points when the epipolar plane is a tangent plane to the surface. At these isolated points the epipolar parametrization can no longer be used to recover the local surface geometry. This paper reviews the epipolar parametrization and shows how the degenerate cases can be used to recover surface geometry and unknown viewer motion from apparent contours of curved surfaces. Practical implementations are outlined.

Visual events' identification of solids of revolution from perspective views

An algorithm is presented for identifying the visual events of solids of revolution from a perspective point of view. The object representation model comprises of these different visual events (or aspects) which can be applied for object identification. Algorithms have been developed to handle general polyhedral objects from either orthographic or perspective views, or to compute the aspect graph of solids of revolution using the orthographic projection. The orthographic view aims to find the aspects of the solid of revolution which are purely object-oriented, the theory of differential geometry can be applied sufficiently to find the aspects without referring to the distance of the viewer. In the real case, the distance effect, that may influence the aspect dramatically when the viewer is close to the object, has to be considered. For instance, to generate the same aspect, the viewer may have to change his viewing directions when he is at a different distance from the object. Firstly, this paper uses the constraint of contour generator to identify the aspects of solids of revolution under perspective projection. The difference between our approach and the others is the distance effect. The present method combines the object-oriented and viewer-oriented aspect finding. Secondly, a method is developed to find the limb termination point of the non-transparent object from a certain perspective viewpoint. Many aspect finding methods are developed for transparent objects (i.e. objects made of glass) of which the projection contour (limb) is continuous. For non-transparent objects, the self-occlusion may induce the limb termination phenomena.

Surface shape from the deformation of apparent contours

The spatiotemporal analysis of deforming silhouettes (apparent contours) is here extended using the mathematics of perspective projections and tools from differential geometry. Analysis of the image motion of a silhouette or apparent contour enables computation of local surface curvature along the corresponding contour generator on the surface, assuming viewer motion is known. To perform the analysis, a spatiotemporal parameterization of image-curve motion is needed, but is underconstrained (a manifestation of the well-known aperture problem). It is shown that an epipolar parameterization is most naturally matched to the recovery of surface curvature.

An EPROM-based programmable contour generator for use in flow cytometry.

An erasable programmable read-only memory (EPROM) contour generator has been fabricated to produce contours for use in flow cytometry. Contours are analog waveforms representing the fluorescence or light-scatter intensity distribution along a cell or object. The generator has particular utility in the development and testing of slit-scan instrumentation and analysis algorithms. Contours are generated without the requirement of specimens or full operation of the flow instrumentation. The generator provides control of contour height, width, offset, and rate. The EPROM may be custom programmed to produce contours for specific test applications or for reproducing "real" contour events. The generator is useful in situations where constant repetitive contours of predetermined characteristics are required.