Abstract: Embodiments of the present invention are directed towards compactly incorporating texture charts into a texture atlas. Texture charts represent three-dimensional mesh segments flattened into two-dimensional shapes. In one embodiment, a texture atlas generating engine is used to generate and evaluate compactness scores of candidate placements for a texture chart. Candidate placements generally refer to the possible locations where a texture chart can be incorporated into a texture atlas. The compactness score can be based on minimizing the distance between a texture chart being incorporated into the texture atlas and the center of mass of previously incorporated texture charts within a texture atlas. In embodiments, an infinity norm can be utilized to compute such a compactness score by outputting an average length of vectors between a texture chart being incorporated into a texture atlas and the texture atlas. Other embodiments may be described and/or claimed.

Abstract: The present disclosure is directed toward systems and methods that facilitate scanning an object (e.g., a three-dimensional object) having custom mesh lines thereon and generating a three-dimensional mesh of the object. For example, a three-dimensional modeling system receives a scan of the object including depth information and a two-dimensional texture map of the object. The three-dimensional modeling system further generates an edge map for the two-dimensional texture map and modifies the edge map to generate a two-dimensional mesh including edges, vertices, and faces that correspond to the custom mesh lines on the object. Based on the two-dimensional mesh and the depth information from the scan, the three-dimensional modeling system generates a three-dimensional model of the object.

Abstract: The present disclosure is directed toward systems and methods that facilitate scanning an object (e.g., a three-dimensional object) having custom mesh lines thereon and generating a three-dimensional mesh of the object. For example, a three-dimensional modeling system receives a scan of the object including depth information and a two-dimensional texture map of the object. The three-dimensional modeling system further generates an edge map for the two-dimensional texture map and modifies the edge map to generate a two-dimensional mesh including edges, vertices, and faces that correspond to the custom mesh lines on the object. Based on the two-dimensional mesh and the depth information from the scan, the three-dimensional modeling system generates a three-dimensional model of the object.

Abstract: The present disclosure is directed toward systems and methods that facilitate scanning an object (e.g., a three-dimensional object) having custom mesh lines thereon and generating a three-dimensional mesh of the object. For example, a three-dimensional modeling system receives a scan of the object including depth information and a two-dimensional texture map of the object. The three-dimensional modeling system further generates an edge map for the two-dimensional texture map and modifies the edge map to generate a two-dimensional mesh including edges, vertices, and faces that correspond to the custom mesh lines on the object. Based on the two-dimensional mesh and the depth information from the scan, the three-dimensional modeling system generates a three-dimensional model of the object.

Abstract: This disclosure relates to methods, non-transitory computer readable media, and systems that use a motion synthesis neural network with a forward kinematics layer to generate a motion sequence for a target skeleton based on an initial motion sequence for an initial skeleton. In certain embodiments, the methods, non-transitory computer readable media, and systems use a motion synthesis neural network comprising an encoder recurrent neural network, a decoder recurrent neural network, and a forward kinematics layer to retarget motion sequences. To train the motion synthesis neural network to retarget such motion sequences, in some implementations, the disclosed methods, non-transitory computer readable media, and systems modify parameters of the motion synthesis neural network based on one or both of an adversarial loss and a cycle consistency loss.

Abstract: This disclosure relates to methods, non-transitory computer readable media, and systems that use a motion synthesis neural network with a forward kinematics layer to generate a motion sequence for a target skeleton based on an initial motion sequence for an initial skeleton. In certain embodiments, the methods, non-transitory computer readable media, and systems use a motion synthesis neural network comprising an encoder recurrent neural network, a decoder recurrent neural network, and a forward kinematics layer to retarget motion sequences. To train the motion synthesis neural network to retarget such motion sequences, in some implementations, the disclosed methods, non-transitory computer readable media, and systems modify parameters of the motion synthesis neural network based on one or both of an adversarial loss and a cycle consistency loss.

Abstract: Methods and systems for generating digital models from objects. In particular, one or more embodiments determine a plurality of correspondences for first and second components of an object. One or more embodiments estimate a joint connecting the first and second components based on the correspondences. One or more embodiments jointly determine a global transformation and one or more joint parameters that map the plurality of components of the object from the first digital scan to the second digital scan. One or more embodiments also updating the correspondences based on the determined global transformation and parameter(s). One or more embodiments re-estimate the joint based on the updated correspondences. One or more embodiments select a candidate joint with a lowest error estimate from a plurality of candidate joints according to determined global transformations and joint parameter(s) for the candidate joints.

Abstract: Embodiments of the present invention are directed towards compactly incorporating texture charts into a texture atlas. Texture charts represent three-dimensional mesh segments flattened into two-dimensional shapes. In one embodiment, a texture atlas generating engine is used to generate and evaluate compactness scores of candidate placements for a texture chart. Candidate placements generally refer to the possible locations where a texture chart can be incorporated into a texture atlas. The compactness score can be based on minimizing the distance between a texture chart being incorporated into the texture atlas and the center of mass of previously incorporated texture charts within a texture atlas. In embodiments, an infinity norm can be utilized to compute such a compactness score by outputting an average length of vectors between a texture chart being incorporated into a texture atlas and the texture atlas. Other embodiments may be described and/or claimed.

Abstract: Embodiments of the present invention are directed towards compactly incorporating texture charts into a texture atlas. Texture charts represent three-dimensional mesh segments flattened into two-dimensional shapes. In one embodiment, a texture atlas generating engine is used to generate and evaluate compactness scores of candidate placements for a texture chart. Candidate placements generally refer to the possible locations where a texture chart can be incorporated into a texture atlas. The compactness score can be based on minimizing the distance between a texture chart being incorporated into the texture atlas and the center of mass of previously incorporated texture charts within a texture atlas. In embodiments, an infinity norm can be utilized to compute such a compactness score by outputting an average length of vectors between a texture chart being incorporated into a texture atlas and the texture atlas. Other embodiments may be described and/or claimed.

Abstract: Certain embodiments involve recognizing combinations of body shape, pose, and clothing in three-dimensional input images. For example, synthetic training images are generated based on user inputs. These synthetic training images depict different training figures with respective combinations of a body pose, a body shape, and a clothing item. A machine learning algorithm is trained to recognize the pose-shape-clothing combinations in the synthetic training images and to generate feature descriptors describing the pose-shape-clothing combinations. The trained machine learning algorithm is outputted for use by an image manipulation application. In one example, an image manipulation application uses a feature descriptor, which is generated by the machine learning algorithm, to match an input figure in an input image to an example image based on a correspondence between a pose-shape-clothing combination of the input figure and a pose-shape-clothing combination of an example figure in the example image.

Abstract: The present disclosure is directed toward systems and methods that facilitate scanning an object (e.g., a three-dimensional object) having custom mesh lines thereon and generating a three-dimensional mesh of the object. For example, a three-dimensional modeling system receives a scan of the object including depth information and a two-dimensional texture map of the object. The three-dimensional modeling system further generates an edge map for the two-dimensional texture map and modifies the edge map to generate a two-dimensional mesh including edges, vertices, and faces that correspond to the custom mesh lines on the object. Based on the two-dimensional mesh and the depth information from the scan, the three-dimensional modeling system generates a three-dimensional model of the object.

Abstract: Certain embodiments involve recognizing combinations of body shape, pose, and clothing in three-dimensional input images. For example, synthetic training images are generated based on user inputs. These synthetic training images depict different training figures with respective combinations of a body pose, a body shape, and a clothing item. A machine learning algorithm is trained to recognize the pose-shape-clothing combinations in the synthetic training images and to generate feature descriptors describing the pose-shape-clothing combinations. The trained machine learning algorithm is outputted for use by an image manipulation application. In one example, an image manipulation application uses a feature descriptor, which is generated by the machine learning algorithm, to match an input figure in an input image to an example image based on a correspondence between a pose-shape-clothing combination of the input figure and a pose-shape-clothing combination of an example figure in the example image.

Abstract: The disclosure describes systems and methods of selecting colors to points in a digital three-dimensional (3D) model representing a scanned object, based on points and color images associated with the 3D model. Certain embodiments involve selecting from the images a patch for each point in the 3D model, and determining a quality of the patches. The selected patches are analyzed to determine an overall score, representing aggregated quality of the patches and an aggregated smoothness indicating variation between patches selected for neighboring points. In some examples, multiple sets of selected patches are analyzed and scored, and the scores are compared to determine a representative patch set that optimizes the quality and the smoothness. Colors are assigned to the points of the digital model based on the representative set of patches.

Abstract: Embodiments of the present invention are directed towards compactly incorporating texture charts into a texture atlas. Texture charts represent three-dimensional mesh segments flattened into two-dimensional shapes. In one embodiment, a texture atlas generating engine is used to generate and evaluate compactness scores of candidate placements for a texture chart. Candidate placements generally refer to the possible locations where a texture chart can be incorporated into a texture atlas. The compactness score can be based on minimizing the distance between a texture chart being incorporated into the texture atlas and the center of mass of previously incorporated texture charts within a texture atlas. In embodiments, an infinity norm can be utilized to compute such a compactness score by outputting an average length of vectors between a texture chart being incorporated into a texture atlas and the texture atlas. Other embodiments may be described and/or claimed.

Abstract: The disclosure describes systems and methods of selecting colors to points in a digital three-dimensional (3D) model representing a scanned object, based on points and color images associated with the 3D model. Certain embodiments involve selecting from the images a patch for each point in the 3D model, and determining a quality of the patches. The selected patches are analyzed to determine an overall score, representing aggregated quality of the patches and an aggregated smoothness indicating variation between patches selected for neighboring points. In some examples, multiple sets of selected patches are analyzed and scored, and the scores are compared to determine a representative patch set that optimizes the quality and the smoothness. Colors are assigned to the points of the digital model based on the representative set of patches.

Abstract: Smoothing images using machine learning is described. In one or more embodiments, a machine learning system is trained using multiple training items. Each training item includes a boundary shape representation and a positional indicator. To generate the training item, a smooth image is downscaled to produce a corresponding blocky image that includes multiple blocks. For a given block, the boundary shape representation encodes a blocky boundary in a neighborhood around the given block. The positional indicator reflects a distance between the given block and a smooth boundary of the smooth image. In one or more embodiments to smooth a blocky image, a boundary shape representation around a selected block is determined. The representation is encoded as a feature vector and applied to the machine learning system to obtain a positional indicator. The positional indicator is used to compute a location of a smooth boundary of a smooth image.

Abstract: Methods and systems for generating digital models from objects. In particular, one or more embodiments determine a plurality of correspondences for first and second components of an object. One or more embodiments estimate a joint connecting the first and second components based on the correspondences. One or more embodiments jointly determine a global transformation and one or more joint parameters that map the plurality of components of the object from the first digital scan to the second digital scan. One or more embodiments also updating the correspondences based on the determined global transformation and parameter(s). One or more embodiments re-estimate the joint based on the updated correspondences. One or more embodiments select a candidate joint with a lowest error estimate from a plurality of candidate joints according to determined global transformations and joint parameter(s) for the candidate joints.

Abstract: Smoothing images using machine learning is described. In one or more embodiments, a machine learning system is trained using multiple training items. Each training item includes a boundary shape representation and a positional indicator. To generate the training item, a smooth image is downscaled to produce a corresponding blocky image that includes multiple blocks. For a given block, the boundary shape representation encodes a blocky boundary in a neighborhood around the given block. The positional indicator reflects a distance between the given block and a smooth boundary of the smooth image. In one or more embodiments to smooth a blocky image, a boundary shape representation around a selected block is determined. The representation is encoded as a feature vector and applied to the machine learning system to obtain a positional indicator. The positional indicator is used to compute a location of a smooth boundary of a smooth image.