Abstract: A non-photorealistic technique is described for transforming an original image into a sketch image. The technique includes: segmenting the original image into plural regions to produce a segmented image, wherein the regions are demarcated by respective boundaries; shrinking a boundary of at least one of the plural regions in the segmented image to produce a boundary-shrunk image; and modifying at least one color in the boundary-shrunk image to produce the sketch image.

Abstract: Techniques for providing localized color transfer are disclosed. In some aspects, a user may select a source region of a source image and a destination region of a destination image. The source region and the destination region may be associated by a designator to create a color transfer pair. A localized color transfer based on the color style of the source region may be implemented to modify the destination region color style. Further aspects may include optimizing the destination image to reduce discontinuities resulting from the color transfer and enabling the user to select regions of the destination image which are not modified by localized color transfer.

Abstract: A face cartooning system is described. In one implementation, the system generates an attractive cartoon face or graphic of a user's facial image. The system extracts facial features separately and applies pixel-based techniques customized to each facial feature. The style of cartoon face achieved resembles the likeness of the user more than cartoons generated by conventional vector-based cartooning techniques. The cartoon faces thus achieved provide an attractive facial appearance and thus have wide applicability in art, gaming, and messaging applications in which a pleasing degree of realism is desirable without exaggerated comedy or caricature.

Abstract: Caricature exaggeration systems, engines, and methods create a drawing of a facial image, compare relationships among facial features in the drawing to corresponding relationships in facial images and corresponding caricatures in a database, and automatically exaggerate at least one relationship among facial features in the drawing based on the facial images and the corresponding caricatures in the database.

Abstract: The techniques and mechanisms described herein are directed to a system for stylizing video, such as interactively transforming video to a cartoon-like style. Briefly stated, the techniques include determining a set of volumetric objects within a video, each volumetric object being a segment. Mean shift video segmentation may be used for this step. With that segmentation information, the technique further includes indicating on a limited number of keyframes of the video how segments should be merged into a semantic region. Finally, a contiguous volume is created by interpolating between keyframes by a mean shift constrained interpolation technique to propagate the semantic regions between keyframes.

Abstract: Mean shift is a nonparametric estimator of density which has been applied to image and video segmentation. Traditional mean shift based segmentation uses a radially symmetric kernel to estimate local density, which is not optimal in view of the often structured nature of image and more particularly video data. The system and method of the invention employs an anisotropic kernel mean shift in which the shape, scale, and orientation of the kernels adapt to the local structure of the image or video. The anisotropic kernel is decomposed to provide handles for modifying the segmentation based on simple heuristics. Experimental results show that the anisotropic kernel mean shift outperforms the original mean shift on image and video segmentation in the following aspects: 1) it gets better results on general images and video in a smoothness sense; 2) the segmented results are more consistent with human visual saliency; and 3) the system and method is robust to initial parameters.

Abstract: The techniques and mechanisms described herein are directed to a system for stylizing video, such as interactively transforming video to a cartoon-like style. Briefly stated, the techniques include determining a set of volumetric objects within a video, each volumetric object being a segment. Mean shift video segmentation may be used for this step. With that segmentation information, the technique further includes indicating on a limited number of keyframes of the video how segments should be merged into a semantic region. Finally, a contiguous volume is created by interpolating between keyframes by a mean shift constrained interpolation technique to propagate the semantic regions between keyframes.

Abstract: Knitwear modeling is disclosed. A macrostructure corresponding to a three-dimensional object is generated, based on a stitch pattern and optionally a color pattern. Yarn microstructure is generated and applied to the macrostructure to yield a knitwear model. The stitch positions of the macrostructure can be perturbed to achieve stitch position irregularities. The fluffiness of the yarn microstructure can be controlled. In an alternative embodiment, a two-dimensional knitwear texture is generated, which can then be mapped to a three-dimensional object to yield a knitwear model.

Abstract: According to embodiments of the invention, an advertisement-generation system generates image-containing advertisements. The advertisement-generation system includes: at least one feature-selection guideline that specifies at least one recommended feature for image-containing advertisements based on advertiser inputs that specify at least one of advertisement-target-audience information, cost information, and advertiser-industry information; an image-clip library from which images are selected for inclusion in the image-containing advertisements; and at least one advertisement template that is based on the at least one feature-selection guideline; wherein the system automatically generates image-containing advertisements that contain one or more suggested colors that are automatically suggested based on one or more colors present on a web page that will host the image-containing advertisement.

Abstract: A non-photorealistic technique is described for transforming an original image into a sketch image. The technique includes: segmenting the original image into plural regions to produce a segmented image, wherein the regions are demarcated by respective boundaries; shrinking a boundary of at least one of the plural regions in the segmented image to produce a boundary-shrunk image; and modifying at least one color in the boundary-shrunk image to produce the sketch image.

Abstract: A process and system for modeling, learning and synthesizing cursive handwriting in a user's personal handwriting style. The handwriting synthesis system and process described herein addresses the problem of learning the personal handwriting style of a user based on limited handwriting samples and producing novel scripts of the same style. The handwriting synthesis process includes segmenting handwriting samples into individual characters using a two-level writer-independent segmentation process, aligning samples of the same character into a common coordinate frame, and learning and modeling the individual character. Synthesis of handwriting is performed by generating individual letters from the models and concatenating the letters using a conditional sampling algorithm. The result is a smooth and fluid connection between letters that successfully mimics the personal handwriting style of a user.

Abstract: Systems and methods for sketching reality are described. In one aspect, a set of vector primitives is identified from a 2-D sketch. In one implementation, the 2-D sketch is hand-drawn by a user. A 2.5D geometry model is automatically generated from the vector primitives. The 2.5D geometry model is automatically rendered and presented to a user. In one implementation, the user provides 2-D sketch-based user inputs to modify one or more of lighting position, lighting direction, lighting intensity, texture, color, and geometry of the presentation.

Abstract: Techniques are described for adaptive changing a displayed foreground color when a conflict between the displayed foreground color and an extracted background color is detected. Upon detection of such a conflict, a new foreground color is selected in accordance with a predetermined legibility criterion. That is, a color pool candidate may be considered to be a viable foreground color if a legibility value for the candidate color in relation to the extracted background color exceeds a predetermined legibility threshold value.

Abstract: A method and system for efficient synthesis of photorealistic free-form knitwear, where a single cross-section of yarn serves as the basic primitive for modeling entire articles of knitwear. This primitive, called the lumislice, describes radiance from a yarn cross-section based on fine-level interactions, including occlusion, shadowing, and multiple scattering, among yarn fibers. By representing yarn as a sequence of identical but rotated cross-sections, the lumislice can effectively propagate local microstructure over arbitrary stitch patterns and knitwear shapes. This framework accommodates varying levels of detail and capitalizes on hardware-assisted transparency blending. To further enhance realism, a technique for generating soft shadows from yarn is also introduced.

Abstract: A system and process for adding a photorealistic rendering of a body of water to a virtual 3D scene or image and creating a video therefrom having interactive water effects. A region of water is added to an image by adding an area depicting the original scene as it would appear if reflected by still body of water. Then, the appearance of the added water region is distorted over a series of image frames in such a way as to simulate how the reflected scene would look if the surface of the water were in motion. The water can have dynamic waves and the user can interact with the water in numbers of ways, including generating ripples on the water surface and creating rain. In addition, these effects can be achieved at full screen resolution with the use of the latest graphics hardware by employing a texture shifting technique.

Abstract: A system and a method for inserting content into a video. A portion of the video is received from a content provider, and data identifying insertion locations in the video is stored. Information describing the video and the insertion locations is received from the content provider. An interface is provided to allow users to view at least a portion of the received information or a portion of the video. A plurality of bids is receiving from the users. Each of the bids relates to insertion of a piece of content into one of the insertion locations. A winning bid is selected from the plurality of bids, and information associated with the winning bid is communicated to the content provider.

Abstract: A method and system for efficient synthesis of photorealistic free-form knitwear, where a single cross-section of yarn serves as the basic primitive for modeling entire articles of knitwear. This primitive, called the lumislice, describes radiance from a yarn cross-section based on fine-level interactions, including occlusion, shadowing, and multiple scattering, among yarn fibers. By representing yarn as a sequence of identical but rotated cross-sections, the lumislice can effectively propagate local microstructure over arbitrary stitch patterns and knitwear shapes. This framework accommodates varying levels of detail and capitalizes on hardware-assisted transparency blending. To further enhance realism, a technique for generating soft shadows from yarn is also introduced.

Abstract: The present invention provides a computer implemented method for placing feathers on a surface. The method includes providing a surface having a plurality of vertices and establishing a growing direction for each of the plurality of vertices on the surface. Feathers are placed on the surface based on the plurality of vertices and the growing direction.

Abstract: A system and process for adding a photorealistic rendering of a body of water to a virtual 3D scene or image and creating a video therefrom having interactive water effects. A region of water is added to an image by adding an area depicting the original scene as it would appear if reflected by still body of water. Then, the appearance of the added water region is distorted over a series of image frames in such a way as to simulate how the reflected scene would look if the surface of the water were in motion. The water can have dynamic waves and the user can interact with the water in numbers of ways, including generating ripples on the water surface and creating rain. In addition, these effects can be achieved at full screen resolution with the use of the latest graphics hardware by employing a texture shifting technique.

Abstract: The present invention provides a systematic method for rendering feathers. In one aspect, a segment of a rachis having a first barb and second barb is generated. The first barb is generated based on a first barb length and the second barb is generated based on a second barb length.