Abstract: A method for reconstructing an image generated from radial trajectory data in frequency or k-space using a GPU. The method includes using a vertex shader of the GPU to transform coordinates of a window aligned with the radial trajectory data and using a pixel shader of the GPU to combine data along the radial trajectory with the coordinate transformed widow to distribute the data along the radial trajectory fed to the pixel shader into cells of a Cartesian coordinate system.

Abstract: A manifold learning technique is applied to the problem of discriminating an object boundary between neighboring pixels/voxels in an image. The manifold learning technique is referred to as locality preserving projections. The application is for multi-channel images, which may include registered images/volumes, a time series of images/volumes, images obtained using different pulse sequences or contrast factors, radar and color photographs.

Abstract: A method and system for segmenting an object in a digital image are disclosed. A user selects at least one foreground pixel or node located within the object and at least one background pixel or node located outside of the object. A random walk algorithm is performed to determine the boundaries of the object in the image. In a first step of the algorithm, a plurality of coefficients is determined. Next, a system of linear equations that include the plurality of coefficients are solved to determine a boundary of the object. The processing is performed by a graphics processing unit. The processing can be performed using the near-Euclidean LUV color space or a Lab color space. It is also preferred to use a Z-buffer in the graphics processing unit during processing. The object, once identified, can be further processed, for example, by being extracted from the image based on the determined boundary.

Abstract: A system and method for segmenting an object in an image using an isoperimetric ratio in a graphics processing unit is disclosed. The object is identified by one or more selected pixels located within the object. The system includes a graphics processing unit and software loadable on the graphics processing unit. The software is operable to find weights for edges, to build a Laplacian matrix and a d vector, to eliminate the row and column corresponding to the one or more selected pixels to determine L0 and d0, to solve x0 in the equation L0x0=d0; and to threshold the potentials x at the value that selects the object having the lowest isoperimetric ratio.

Abstract: A method for assigning stiffness values to voxels of an inputted discretized image includes assigning stiffness values according to intensity values of voxels of the image, wherein the inputted discretized image comprises the intensity values and refining the stiffness values assigned to voxels using a segmentation.

Abstract: A method for on-the-fly segmentations for image deformations, includes: selecting an object in an image that is displayed on a computer screen; immediately after selecting the object, automatically segmenting the object to find a boundary of the object and marking an area inside the boundary undeformable and an area outside the boundary deformable; and dragging the object to a desired location in the image while the area inside the boundary remains stiff and the area outside the boundary deforms.

Abstract: A method for deforming an image using a physical model includes inputting the image, discretizing the image into a grid, applying forces to the grid upon receiving an input command for dragging the grid, updating a mesh deformation of the grid according to the forces, and wrapping the image according to a deformed mesh to produce a deformed image, wherein a wrapped image is displayed on an display device.

Abstract: A method for reconstructing an object subject to cone beam passing through the object and generating projection images of the object using a graphic processing unit (GPU). The method applies a non-linear curvature-based smoothing filter to the projection images. The method applies a high-pass filter to the curvature-smoothed images. The method backprojects the projection images into an output volume using voxel driven backprojection. The method removes cupping artifacts from the output volume convolving every slice with a Butterworth kernel and adding the result to the slice weighted by a scaling factor.

Abstract: A method for reconstructing an image generated from radial trajectory data in frequency or k-space using a GPU. The method includes using a vertex shader of the GPU to transform coordinates of a window aligned with the radial trajectory data and using a pixel shader of the GPU to combine data along the radial trajectory with the coordinate transformed widow to distribute the data along the radial trajectory fed to the pixel shader into cells of a Cartesian coordinate system.

Abstract: A GPU accelerated volume rendering method from flat textures (also known as texture atlas) is disclosed. The method is not restricted to a specific rendering technique. The method is fast as it requires no reordering or copying passes. An image processing system is also provided. Addressing of two dimensional flat textures is accomplished based on the viewing direction. A first addressing scheme is used for the x direction, a second addressing scheme is used for the y direction and a third addressing scheme is used for the z direction, where x,y, and z refer to the axes of Cartesian coordinate system.

Abstract: A manifold learning technique is applied to the problem of discriminating an object boundary between neighboring pixels/voxels in an image. The manifold learning technique is referred to as locality preserving projections. The application is for multi-channel images, which may include registered images/volumes, a time series of images/volumes, images obtained using different pulse sequences or contrast factors, radar and color photographs.

Abstract: A system and method for segmenting an object in an image using an isoperimetric ratio in a graphics processing unit is disclosed. The object is identified by one or more selected pixels located within the object. The system includes a graphics processing unit and software loadable on the graphics processing unit. The software is operable to find weights for edges, to build a Laplacian matrix and a d vector, to eliminate the row and column corresponding to the one or more selected pixels to determine L0 and d0, to solve x0 in the equation L0x0=d0; and to threshold the potentials x at the value that selects the object having the lowest isoperimetric ratio.

Abstract: A method and system for segmenting an object in a digital image are disclosed. A user selects at least one foreground pixel or node located within the object and at least one background pixel or node located outside of the object. A random walk algorithm is performed to determine the boundaries of the object in the image. In a first step of the algorithm, a plurality of coefficients is determined. Next, a system of linear equations that include the plurality of coefficients are solved to determine a boundary of the object. The processing is performed by a graphics processing unit. The processing can be performed using the near-Euclidean LUV color space or a Lab color space. It is also preferred to use a Z-buffer in the graphics processing unit during processing. The object, once identified, can be further processed, for example, by being extracted from the image based on the determined boundary.