Abstract: Processing images of a scene taken with different focus includes, for each of at least some of the images, determining at least one portion of the image having a predetermined characteristic. A representation of the scene is generated that includes an array of elements, where each of at least some of the elements corresponds to a determined portion of at least one of the images. An element of the array is based on a correspondence between a distance associated with a determined portion of at least one image and a focused distance associated with at least one image. The generated representation is stored or provided as output.

Abstract: A method includes receiving into a computer processor a building control subsystem design drawing, and identifying a plurality of objects in the building control subsystem design drawing by comparing the objects to a template of objects. The template of objects includes one or more of a representation of a shape, a color, and a texture. A physical relationship among the plurality of objects is determined, and a three dimensional representation of the plurality of objects is retrieved from a three dimensional device library. A three dimensional building control subsystem graphic is generated by mapping the three dimensional representation of the plurality of objects into a three dimensional user interface as a function of the physical relationship. The three dimensional user interface is animated and interactive to monitor and control a building subsystem.

Abstract: Generating three-dimensional information can include accessing multiple different images of an object taken by one or more cameras; selecting one of the accessed images as a reference image; identifying corresponding features between the reference image and one or more different ones of the accessed images; determining first camera pose information for each accessed image based on one or more of the corresponding features, each first camera pose information indicative of a relationship between an imaging device and the object; determining a first three-dimensional structure of the object based on first camera pose information of two of the accessed images; and generating a second three-dimensional structure and a second camera pose information for each accessed image based on the first three-dimensional structure and the first camera pose information for each accessed image.

Abstract: A method and system for producing an image includes creating an image template having a customizable region and extracting a 2D object from a source image. The 2D object is transformed using a 3D model in response to the customizable region, and the transformed 2D object is merged into the image template to create a customized image.

Abstract: A self-localization device and a method thereof are provided. The self-localization device has a movable carrier, a first laser image-taking device and a processor. The movable carrier can be moved and rotated on a plan. During the motion of the movable carrier, the first laser image-taking device disposed on the movable carrier acquires an i-th lot point data in the space at a time point ti, where i is one index number from 1 to n, and n is an integer. The processor controls the first laser image-taking device, and receives coordinates of the i-th lot point data. The processor executes a K-D tree algorithm to perform a comparison and merge process between the first and the i-th lots point data, so as to establish a two dimensional profile.

Abstract: Arrangements are disclosed herein that can capture image data reflecting real world objects and/or interactions and convert this data into a format that is usable in a virtual universe (VU). In one embodiment, a method can include receiving sets of image data that capture an object or a setting from different viewing angles. The image can be in multiple data sets where the sets have at least one common feature. The image data can be analyzed using the common features in the image data and a spatial map can be created. The spatial map can be modified (if required) such that a VU engine can use the modified data to provide a VU that has a representation of the real world object and/or interaction. This allows a participant to customize objects and/or activities in the VU, and provide a unique “persona” for their avatar and their VU.

Abstract: In a cost-efficient method and arrangement for 3D digitization of bodies and body parts, which produces dense and exact spatial coordinates despite imprecise optics and mechanics, the body to be digitized is placed on a photogrammetrically marked surface, a photogrammetrically marked band is fitted to the body or body part to be digitized, and a triangulation arrangement comprised of a camera and a light pattern projector is moved on a path around the body. By a photogrammetric evaluation of the photogrammetric marks of the surface and the band situated in the image field of the camera, and of the light traces of the light projector on the marked surface and the marked band, all unknown internal and external parameters of the triangulation arrangement are determined, and the absolute spatial coordinates of the body or body part are established from the light traces on the non-marked body with high point density and high precision without any separate calibration methods.

Abstract: A method for ascertaining the axis of rotation of a vehicle wheel in which a light pattern is projected at least onto the wheel during the rotation of the wheel and the light pattern reflected from the wheel is detected by a calibrated imaging sensor system and analyzed in an analyzer device. Accurate and robust measurement of the axis of rotation and, optionally, of the axis and wheel geometry, in particular when the vehicle is passing by, is achieved in that a 3D point cloud with respect to the wheel is determined in the analysis and a parametric surface model of the wheel is adapted thereto; normal vectors of the wheel are calculated for different rotational positions of the wheel for obtaining the axes of rotation; and the axis of rotation vector is calculated as the axis of rotation from the spatial movement of the normal vector of the wheel.

Abstract: In one embodiment, a system and method for modeling a three-dimensional object includes capturing two-dimensional images of the object from multiple different viewpoints to obtain multiple views of the object, estimating slices of the object that lie in parallel planes that cut through the object, and computing a surface of the object from the estimated slices.

Abstract: In a device and associated method for reconstruction and visualization of projection data, projection data are stored per slice and are subjected to an image reconstruction procedure in parallel within arbitrary slice planes in a processor-controlled filtering process that is executed n times, wherein volume data that are created can already be made available (loaded) for a direct visualization.

Abstract: A method for synthetic reconstruction of objects includes: extracting criteria from a knowledge base; extracting, from sensed signals filtered by the criteria, weak signals; extracting, from the weak signals, weak signals of interest; removing noise from and amplifying the weak signals of interest and obtaining useful weak signals; identifying useful direct information, from useful weak signals filtered by the criteria and supplying optimum criteria; reconstructing, using the useful direct information, information of interest; reconstructing, using the information of interest, useful information and supplying optimum criteria; reconstructing, based on the useful information, three-dimensional information, supplying a recognition state file and supplying the optimum criteria; and updating the criteria with the optimum criteria.

Abstract: The claimed subject matter provides a system and/or a method that facilitates simulating a portion 2-dimensional (2D) data for implementation within a 3-dimensional (3D) virtual environment. A 3D virtual environment can enable a 3D exploration of a 3D image constructed from a collection of two or more 2D images, the 3D image is constructed by combining the two or more 2D images based upon a respective image perspective. An analyzer can evaluate the collection of two or more 2D images to identify a portion of the 3D image that is unrepresented by the combined two or more 2D images. A synthetic view generator can create a simulated synthetic view for the portion of 3D image that is unrepresented, the simulated synthetic view replicates a 2D image with a respective image perspective for the unrepresented portion of 3D image.

Abstract: The process and method for generating a 3D model from a set of 2D drawings is described herein. Traditionally, many structural components (objects) are communicated through a series of 2D drawings, wherein each drawing describes the components that are visible in a user-selected view direction. No machine-readable information in the drawings define a relationship between the drawings developed from various view directions or the objects' locations in 3D space. Considerable human effort and intervention is required to place objects defined in the 2D drawings into 3D space. With the ability to provide information in each drawing defining a relationship with the other drawings as well as its place in 3D space, the objects defined in 2D drawings can self-assemble in 3D space, thereby reducing a substantial amount of required human effort.

Abstract: A method for associating a three-dimensional surface representing a real object and a three-dimensional reference surface, said reference surface being represented by a set of reference points, the method comprising: obtaining a set of real points representing the real surface, determining the normal vector of each point of said obtained set of real points, selecting, among the set of real points, control points according to the determined normal vector by converting the set of real points to a bi-dimensional space of normal vectors, generating sets of points having similar normal vector among the points of the set of real points and selecting, for each set of points with similar normal vector, one point that is a control point of the real surface, determining correspondence points close to the set of reference points that are determined to correspond to the control points of the real surface, and determining the motion that minimizes the distances between the control points of the real surface and the corres

Abstract: Window based matching is used for determining a depth map from images obtained from different orientations. A set of fixed matching windows is used for points of the image for which the depth is to be determined. The set of matching windows covers a footprint of pixels around the point of the image, and the average number (O) of matching windows that a pixel of the footprint (FP) belongs to is less than one plus the number of pixels in the footprint divided by 15 (O<FP/15+1), preferably less than one plus the number of pixels in the footprint divided by 25 (O<FP/25+1).

Abstract: An image processing apparatus includes a memory unit which stores data of a first projection image and data of a second projection image, which are associated with the same object and are captured in different imaging directions, a display unit which displays the data of the first projection image and the data of the second projection image, a designation operation unit which is configured to designate a plurality of points on the displayed first and second projection images, and an operation supporting unit which generates operation supporting information for supporting an operation of designating, by the designation operation unit, the plurality of points on the second image, which anatomically correspond to the plurality of points designated on the first projection image.

Abstract: A medical image processing apparatus which performs image processing on a medical input image, includes: a pixel extraction portion for extracting an inappropriate pixel satisfying a predetermined condition in the input image, a replacing pixel information calculation portion for calculating replacing information with which pixel information of the inappropriate pixel is to be replaced, based on pixel information of a predetermined region including the extracted inappropriate pixel or adjacent to the extracted inappropriate pixel and a replaced image generation portion for replacing the pixel information of the inappropriate pixel in the input image with the replacing information and generating a replaced image.

Abstract: A method for processing data includes identifying a time signature of an infra-red (IR) beacon. Image data associated with the IR beacon is identified using the time signature.

Abstract: A shape measuring apparatus that measures a three-dimensional shape of a measuring target has a lighting device that irradiates the measuring target placed on a stage with light, an imaging device that takes an image of the measuring target, a shape calculating device that calculates orientations of normals at a plurality of points on a surface of the measuring target from an image, the image being obtained by performing imaging with the imaging device while the lighting device irradiates the measuring target with the light, the shape calculating device calculating the three-dimensional shape of the surface of the measuring target from the calculation result of the orientations of the normals, a ranging device that measures a distance from a predetermined reference position with respect to at least one point on the surface of the measuring target, and a determination device that determines a spatial position of the three-dimensional shape of the surface of the measuring target, the three-dimensional shape bei

Abstract: A system and method of auto-exposure control is provided for image acquisition hardware using three dimensional information to identify a region(s) of interest within an acquired 2D image or images upon which to apply traditional auto-exposure techniques. By performing auto-exposure analysis over the region of interest, the acquisition property settings can be assigned such that the light levels within the region of interest fall within the linear range, producing sufficient grayscale information for identifying particular objects and profiles in subsequently acquired images. For example, in a machine vision application that detects people passing through a doorway, the region of interest can be the portion of the 2D image that generated 3D features of a head and shoulders profile within a 3D model of the doorway scene. With higher quality images, more accurate detection of people candidates within the monitored scene results.

Abstract: A method for preparing a clinical restraint for a subject, the method comprises scanning the portion of the subject to be restrained to produce a 3D image data set, generating a three dimensional replica of the portion of the subject from the 3D image data set and preparing a clinical restraint using the three dimensional replica. A scanning system for generating the 3D image data set comprising one or more projectors (6) and one or more cameras (8) in combination with an image processing device is also disclosed. The imaging aspects are also applied in the monitoring of the treatment of a patient, the manufacture and fitting of medical items, such as compression hosiery and the like, as well as in the fitting of garments and items of clothing.

Abstract: The present invention provides a system for generating images of multi-views.

Abstract: A singular value decomposition method according to the present invention is a method for performing a singular value decomposition on an arbitrary matrix A using a computer, the method including the steps of: performing an upper bidiagonalization on the matrix A so as to obtain an upper bidiagonal matrix B of the matrix A; obtaining at least one singular value ? of the matrix B as singular values of the matrix A; and obtaining a singular vector of the matrix A for the ?. The step of obtaining a singular vector of the matrix A includes a step of performing a Twisted decomposition on a matrix BTB??2I (where I is a unit matrix) by using a Miura inverse transformation, an sdLVvs transformation, an rdLVvs transformation and a Miura transformation so as to diagonalize a matrix BTB.

Abstract: A method for a computer system including receiving a file comprising textures including a first and a second texture, and metadata, wherein the first texture need not have a predetermined geometric relationship to the second texture, wherein the metadata includes identifiers associated with textures and includes adjacency data, associating the first texture with a first location on an object in response to an identifier associated with the first texture, associating the second texture with a second location on the object in response to an identifier associated with the second texture, determining an edge of the first texture is adjacent to an edge of the second texture in response to the adjacency data, and performing a rendering operation with respect to the first and the second surface on the object to determine rendering data in response to the first texture and to the second texture.

Abstract: An object of the present invention is to enable performing height recognition processing by setting a height of an arbitrary plane to zero for convenience of the recognition processing. A parameter for three-dimensional measurement is calculated and registered through calibration and, thereafter, an image pickup with a stereo camera is performed on a plane desired to be recognized as having a height of zero in actual recognition processing. Further, three-dimensional measurement using the registered parameter is performed on characteristic patterns (marks m1, m2 and m3) included in this plane. Three or more three-dimensional coordinates are obtained through this measurement and, then, a calculation equation expressing a plane including these coordinates is derived.

Abstract: An image may be dehazed using a three-dimensional reference model. In an example embodiment, a device-implemented method for dehazing includes acts of registering, estimating, and producing. An image that includes haze is registered to a reference model. A haze curve is estimated for the image based on a relationship between colors in the image and colors and depths of the reference model. A dehazed image is produced by using the estimated haze curve to reduce the haze of the image.

Abstract: Embodiments of the invention disclose a system and a method for determining a disparity search range for a current stereo image of a scene based on a set of stereo images of the scene, comprising steps of: selecting a subset of stereo images from the set of stereo images, the subset includes the current stereo image and at least one neighboring stereo image, wherein the neighboring stereo image is temporally-neighboring to the current stereo image; determining a disparity histogram for each stereo image in the subset of stereo images to form a set of disparity histograms; determining a weighted disparity histogram as a weighted sum of the disparity histograms in the set of disparity histograms; and determining the disparity search range from the weighted disparity histogram.

Abstract: A method is disclosed for screening color separations of a lenticular image with a lenticular frequency of the lenticular lenses needed for viewing the lenticular image. An amplitude-modulated halftone image is calculated for each of the color separations at a screen angle and at a screen frequency, the tangent of the screen angle being a rational number. For a specific color separation, a screen angle relating to the direction perpendicular to the image strips of the lenticular image is defined, a pair of whole numbers whose ratio is equal to the tangent of the screen angle is determined, and the screen frequency of the color separation is calculated as the product of the lenticular frequency and the square root of the sum of the squares of the two whole numbers.

Abstract: There is provided an image printer having a display unit, in which a predetermined image can be set as a background picture of the display unit. The image printer includes an image processing unit processing an image selected by a user in accordance with a predetermined image processing technique so as to set the selected image as a background picture of the display unit; and a setting unit setting the processed image as the background picture of the display unit, wherein the display unit displays the processed image.

Abstract: Site modeling using image data fusion. Geometric shapes are generated to represent portions of one or more structures based on digital height data and a two-dimensional segmentation of portions of the one or more structures is generated based on three-dimensional line segments and digital height data. A labeled segmentation of the one or more structures is generated based on the geometric shapes and the two-dimensional segmentation. A three-dimensional model of the one or more structures is generated based on the labeled segmentation.

Abstract: In one embodiment, a method comprises determining an initial set of feature correspondences for a pair of images of a three dimensional (3D) space; identifying one or more first regions in one of the pair of images and one or more second regions in the other one of the pair of images. The method comprises identifying additional feature correspondences by matching first regions and second regions in the pair of images and generating a motion of a camera in the 3D space responsive to the initial set of feature correspondences and the additional feature correspondences, wherein the camera captured the images. Each of the first regions and the second regions potentially maps to a plane in the 3D space corresponding to the images. Additionally, each of the first regions and the second regions includes at least one feature in the initial set of feature correspondences.

Abstract: A scanning apparatus includes a scanning head that scans an image of a document positioned on a stage glass, but varying in its distance therefrom. A boundary line is detected to determine an amount of skew therein. A skew line is compared with an established reference line, and a correction factor is calculated based on the result of the comparison. The original image processed to map the boundary line to the reference line and image data inside of the boundary line is similarly mapped based on the calculated correction factor. Beneficially, skews of an image produced when a thick book is scanned can be corrected simply.

Abstract: A method of rendering a multi-view image comprising a first output image and a second output image on basis of an input image (102) is disclosed. The method comprises: creating a modulation image (100) comprising irregular shaped objects (106-112); modulating pixel values of a portion of the input image (102) on basis of further pixel values of the modulation image (100), resulting into an intermediate image (104); and generating the multi-view image by means of warping the intermediate image on basis of the disparity data.

Abstract: A method for reconstructing a 3D shape model of an object by using multi-view image information, includes: inputting multi-view images obtained by photographing the object from multiple viewpoints in a voxel space, and extracting silhouette information and color information of the multi-view images; reconstructing visual hulls by silhouette intersection using the silhouette information; and approximating polygons of cross-sections of the visual hulls to a natural geometric shape of the object by using the color information. Further, the method includes expressing a 3D geometric shape of the object by connecting the approximated polygons to create a mesh structure; extracting color textures of a surface of the object by projecting meshes of the mesh structure to the multi-view image; and creating a 3D shape model by modeling natural shape information and surface color information of the object.

Abstract: The present invention provides a true-scale, coordinate-matched, linked in real-time, dual three-dimensional/two-dimensional visual display/viewer. The display simultaneously shows a 3D digital image and an associated 2D digital image of a selected drawing. The display of the present invention allows a user to visualize an asset's location, surrounding environment and hazards and true scale structural details for interior or external structural scenes. Using the display and associated tools, the user can obtain real-time information of an environment, true-scale measurement, plan ingress/egress paths, shortest paths between points and the number of doorways in a structure and track objects within the displayed environment. The intelligence gained using the tools and 3D/2D display may be used and further manipulated by a single user or may be distributed to other users.

Abstract: A computer-implemented method for extracting boundary elements of an object includes receiving measurement data of a point cloud of the object and a maximum edge length of a triangle to form a triangular mesh surface of the point cloud, constructing the triangular mesh surface, and extracting boundary points from the triangular mesh surface. The method further includes grouping the boundary points according to a vector relationship between one boundary point and each of the other boundary points, fitting each group of boundary points to an appropriate boundary characteristic. The appropriate boundary characteristic may be output to a storage system.

Abstract: The present invention is a system that grids original data, maps the data at the grid locations to height values at corresponding landscape image pixel locations and renders the landscape pixels into a three-dimensional (3D) landscape image. The landscape pixels can have arbitrary shapes and can be augmented with additional 3D information from the original data, such as an offset providing additional information, or generated from processing of the original data, such as to alert when a threshold is exceeded, or added for other purposes such as to point out a feature. The pixels can also convey additional information from the original data using other pixel characteristics such as texture, color, transparency, etc.

Abstract: A method for constructing a georeferencing-enabled camera model and its deployment in georeferencing targets located in video sequences due to a single video camera. A video surveillance camera is modeled by a collection of rays converging at a virtual camera point and the retina resolution cell coordinates associated with those rays wherein the ray equations are first established, in the course of a calibration process, for a given camera view, with the aid of other views of the same video surveillance camera, as necessary, and using such a model for mapping image coordinates to terrain coordinates and vice versa in the intended view or its adaptation for use in other views of the same video surveillance camera.

Abstract: Disclosed are embodiments of systems and methods for synthesizing a detailed depth map from a video image. In embodiments, the motion vectors decoded from a video stream may be classified into groups by the application of K-Model clustering techniques based on an affine model. In embodiments, a coarse depth map of the image pixels may be generated using the image segmented according to the motion vector clusters. In embodiments, high resolution gradient maps of the image may be generated using the coarse depth map as well as edge information from the image. In embodiments, a surface reconstruction algorithm, such as the Frankot-Chellappa algorithm, may be applied to the high resolution gradient maps to synthesize a detailed depth map of the image. A detailed depth map of an image may be used to render a three-dimensional surface, for example.

Abstract: Three-dimensional imaging techniques are used for a visualization method and apparatus. In a preferred embodiment, terrain data is displayed as a series of pixels—areas of terrain elevation data. Individual pixels are analyzed to determine whether they are locally smooth or “warpable” relative to their surrounding neighbor pixels. Those pixels that are locally relatively “smooth,” i.e., those satisfying a given set of criteria, are joined with adjacent neighbor pixels by a process referred to herein as “warping” to create “smooth,” gap-free surfaces. A preferred embodiment includes drawing or generating lines between the centers of two pairs of adjacent pixels to determine a slopes m1 and m2 respectively. The slopes m1 and m2 are then analyzed using the following equations/determinations: |m1??mmax; |m2??mmax; and |m1?m2|??max; i.e.

Abstract: The claimed subject matter provides a system and/or a method that facilitates displaying relevant advertisements to a user. A display engine can browse a portion of image data during a browsing session. An evaluator can identify a context related to two or more concurrent and on-going browsing sessions. An ad selector can locate an ad from a data store based on the identified context and seamlessly incorporate and display the ad into at least one of the browsing sessions.

Abstract: Printed material for stereoscopic viewing is fabricated by creating images for the left eye and the right eye IL1 and IR1 for stereoscopic viewing, performing correction processing to remove perspective of images in the IL1 and IR1 at a base surface, creating images for the left eye and the right eye IL2 and IR2, and performing anaglyph processing or the like, based on the IL2 and IR2. An image IL for the left eye is created by rendering a projection of each point of a graphic object onto a base surface BS in a projection direction linking a viewpoint position for the left eye VPL to each point of the graphic object OB, and an IMAGE IR FOR THE RIGHT EYE is created by rendering a projection of each point of the graphic object onto the base surface BS in a projection direction linking a viewpoint position for the right eye VPL to each point of the graphic object OB. Printed material for stereoscopic viewing is fabricated by anaglyph processing or the like, based on the IL and IR.

Abstract: A computer implemented method and system for constructing a three dimensional (3D) tomographic image from an object's two dimensional (2D) panoramic image are provided. A first geometrical attribute set in a first coordinate system is assigned to one or more focal troughs. A second geometrical attribute set in a second coordinate system is assigned to the 2D panoramic image. Second geometrical attributes are correlated with first geometrical attributes for reconstructing the 2D panoramic image in multiple dimensions. Multiple defocused elements of the object are determined along the 2D panoramic image's horizontal dimension. A transverse dimension is determined for the reconstructed panoramic image by mapping the defocused elements to a translation along the transverse dimension in the first coordinate system on either side of the center of the focal troughs. The multiple dimensions of the reconstructed panoramic image are transformed into an orthogonal coordinate system to generate the 3D tomographic image.

Abstract: Apparatus for processing of a LiDAR point cloud of a ground scan, comprises: a point cloud input for receiving said LiDAR point cloud, a ground filter for filtering out points that belong to the ground from said point cloud, thereby to generate an elevation map showing features extending from the ground, an automatic feature search and recognition unit associated with said three dimensional graphical engine for searching said elevation map of said three-dimensional model to identify features therein and to replace points associated with said feature with a virtual object representing said feature, thereby to provide objects within said data; and a three-dimensional graphical renderer supporting three-dimensional graphics, to generate a three-dimensional rendering of said ground scan.

Abstract: Provided is a camera self-calibration method that calculates a focal length of a fixed zoom lens camera from a correspondence point position between images. In the method, a cost function, which is a function of a focal length, is defined, and a focal length that minimizes the defined cost function is obtained to obtain a focal length that allows 3D recovery results of correspondence points calculated from all image pairs coincide with one another. Therefore, reliability of the calculated focal length can be easily verified, and the focal length of the camera can be stably calculated even when the position of input correspondence point is inaccurately given.

Abstract: An optical flux is radiated, which is focused at a measurement point in a specimen space, and a transmitted light amount is measured. A minute light absorption amount is measured from a transmitted light signal and a reference signal. While three-dimensionally scanning, a three-dimensional map in which light absorption amounts are represented by voxels (volume cells) is obtained. On this three-dimensional map, deconvolution processing with a light intensity distribution image in the vicinity of the measurement point being a convolution kernel is performed, so as to obtain a three-dimensional image of a specimen that is almost transparent in a non-dyed state.

Abstract: A method for determining 3D distances on a 2D pixelized image of a part or object includes acquiring a real 2D pixelized image of the object, creating a simulated image of the object using the 3D CAD model and the 2D pixelized image, determining a specified cost function comparing the simulated image with the real 2D pixilated image and repositioning the simulated image in accordance with iterated adjustments of a relative position between the CAD model and the 2D pixilated image to change the simulated image until the specified cost function is below a specified value. Then, the workstation is used to generate a 3D distance scale matrix using the repositioned simulated image, and to measure and display distances between selected pixels on a surface of the real image using 2D distances on the 2D pixelized image of the object and the 3D distance scale matrix.

Abstract: According to one embodiment, an image conversion apparatus includes a 3D conversion instruction module, a determination module, and a converter. The 3D conversion instruction module is configured to instruct execution of a 3D conversion required to convert an input image into a 3D image. The determination module is configured to determine validity or invalidity of the 3D conversion instruction based on a type of the input image. The converter is configured to convert, based on validity determination of the 3D conversion instruction, the input image into the 3D image in response to the 3D conversion instruction.

Abstract: Creating a 3D face reconstruction model using a single 2D image and a generic facial depth map that provides depth information. In one example, the generic facial depth map is selected based on gender and ethnicity/race. In one embodiment, a set of facial features of the 2D image is mapped to create a facial-feature map, and a 2D mesh is created using the map. The same set of facial features is also mapped onto a generic facial depth map, and a 3D mesh is created therefrom. The 2D image is then warped by transposing depth information from the 3D mesh of the generic facial depth map onto the 2D mesh of the 2D image so as to create a reconstructed 3D model of the face. The reconstructed 3D model can be used, for example, to create one or more synthetic off-angle-pose images of the subject of the original 2D image.

Abstract: A dynamic wide angle image viewing technique is presented which provides a way to view a wide-angle image while zooming between a wide angle view and a narrower angle view that employs both perspective and non-perspective projection models. In general, this involves first establishing the field of view for a view of the wide angle image that is to be displayed. The view is then rendered and displayed based on the established field of view, such that the projection transitions between a perspective projection associated with narrower angle views and a non-perspective projection (e.g., cylindrical, spherical or some other parameterization) associated with wider-angle views.