Abstract: A computing system is provided for generating a synthesized image using subject-driven text-to-image generation. The computing system includes a processing circuitry and memory storing instructions that, when executed, cause the processing circuitry to receive one or more input images of a subject and a user prompt, generate a mosaic of two or more component images based on the one or more input images and the user prompt, generate a mask to mark a placeholder area in the mosaic to be inpainted, use a generative model to inpaint the placeholder area, extract the inpainted placeholder area to generate an extracted image of the subject, and output the extracted image of the subject as the synthesized image.

Abstract: The present disclosure describes techniques for automatically identifying a checkpoint of a machine learning model for deployment. A plurality of checkpoints are generated during training the machine learning model. The machine learning model is trained on a set of training images. A plurality of subject images is generated by each of the plurality of checkpoints. Subject similarity and global difference between images in each pair of images are computed. Each pair of images comprises one of the set of training images and one of the plurality of subject images generated by each of the plurality of checkpoints. Image generation qualities of the plurality of checkpoints are evaluated based on the subject similarity and the global difference. The checkpoint of the machine learning model for deployment is automatically identified based on the evaluated image generation qualities.

Abstract: The present disclosure describes systems, non-transitory computer-readable media, and methods for utilizing a machine learning model trained to determine subtle pose differentiations to analyze a repository of captured digital images of a particular user to automatically capture digital images portraying the user. For example, the disclosed systems can utilize a convolutional neural network to determine a pose/facial expression similarity metric between a sample digital image from a camera viewfinder stream of a client device and one or more previously captured digital images portraying the user. The disclosed systems can determine that the similarity metric satisfies a similarity threshold, and automatically capture a digital image utilizing a camera device of the client device. Thus, the disclosed systems can automatically and efficiently capture digital images, such as selfies, that accurately match previous digital images portraying a variety of unique facial expressions specific to individual users.

Abstract: The present disclosure describes systems, non-transitory computer-readable media, and methods for utilizing a machine learning model trained to determine subtle pose differentiations to analyze a repository of captured digital images of a particular user to automatically capture digital images portraying the user. For example, the disclosed systems can utilize a convolutional neural network to determine a pose/facial expression similarity metric between a sample digital image from a camera viewfinder stream of a client device and one or more previously captured digital images portraying the user. The disclosed systems can determine that the similarity metric satisfies a similarity threshold, and automatically capture a digital image utilizing a camera device of the client device. Thus, the disclosed systems can automatically and efficiently capture digital images, such as selfies, that accurately match previous digital images portraying a variety of unique facial expressions specific to individual users.

Abstract: Embodiments are disclosed for generating lens blur effects. The disclosed systems and methods comprise receiving a request to apply a lens blur effect to an image, the request identifying an input image and a first disparity map, generating a plurality of disparity maps and a plurality of distance maps based on the first disparity map, splatting influences of pixels of the input image using a plurality of reshaped kernel gradients, gathering aggregations of the splatted influences, and determining a lens blur for a first pixel of the input image in an output image based on the gathered aggregations of the splatted influences.

Abstract: The present disclosure describes systems, non-transitory computer-readable media, and methods for utilizing a machine learning model trained to determine subtle pose differentiations to analyze a repository of captured digital images of a particular user to automatically capture digital images portraying the user. For example, the disclosed systems can utilize a convolutional neural network to determine a pose/facial expression similarity metric between a sample digital image from a camera viewfinder stream of a client device and one or more previously captured digital images portraying the user. The disclosed systems can determine that the similarity metric satisfies a similarity threshold, and automatically capture a digital image utilizing a camera device of the client device. Thus, the disclosed systems can automatically and efficiently capture digital images, such as selfies, that accurately match previous digital images portraying a variety of unique facial expressions specific to individual users.

Abstract: The present disclosure relates to systems, methods, and non-transitory computer-readable media that generate refined segmentation masks for digital visual media items. For example, in one or more embodiments, the disclosed systems utilize a segmentation refinement neural network to generate an initial segmentation mask for a digital visual media item. The disclosed systems further utilize the segmentation refinement neural network to generate one or more refined segmentation masks based on uncertainly classified pixels identified from the initial segmentation mask. To illustrate, in some implementations, the disclosed systems utilize the segmentation refinement neural network to redetermine whether a set of uncertain pixels corresponds to one or more objects depicted in the digital visual media item based on low-level (e.g., local) feature values extracted from feature maps generated for the digital visual media item.

Abstract: The present disclosure describes systems, non-transitory computer-readable media, and methods for utilizing a machine learning model trained to determine subtle pose differentiations to analyze a repository of captured digital images of a particular user to automatically capture digital images portraying the user. For example, the disclosed systems can utilize a convolutional neural network to determine a pose/facial expression similarity metric between a sample digital image from a camera viewfinder stream of a client device and one or more previously captured digital images portraying the user. The disclosed systems can determine that the similarity metric satisfies a similarity threshold, and automatically capture a digital image utilizing a camera device of the client device. Thus, the disclosed systems can automatically and efficiently capture digital images, such as selfies, that accurately match previous digital images portraying a variety of unique facial expressions specific to individual users.

Abstract: Embodiments are disclosed for generating lens blur effects. The disclosed systems and methods comprise receiving a request to apply a lens blur effect to an image, the request identifying an input image and a first disparity map, generating a plurality of disparity maps and a plurality of distance maps based on the first disparity map, splatting influences of pixels of the input image using a plurality of reshaped kernel gradients, gathering aggregations of the splatted influences, and determining a lens blur for a first pixel of the input image in an output image based on the gathered aggregations of the splatted influences.

Abstract: Embodiments are disclosed for generating lens blur effects. The disclosed systems and methods comprise receiving a request to apply a lens blur effect to an image, the request identifying an input image and a first disparity map, generating a plurality of disparity maps and a plurality of distance maps based on the first disparity map, splatting influences of pixels of the input image using a plurality of reshaped kernel gradients, gathering aggregations of the splatted influences, and determining a lens blur for a first pixel of the input image in an output image based on the gathered aggregations of the splatted influences.

Abstract: The present disclosure relates to systems, methods, and non-transitory computer-readable media that generate refined segmentation masks for digital visual media items. For example, in one or more embodiments, the disclosed systems utilize a segmentation refinement neural network to generate an initial segmentation mask for a digital visual media item. The disclosed systems further utilize the segmentation refinement neural network to generate one or more refined segmentation masks based on uncertainly classified pixels identified from the initial segmentation mask. To illustrate, in some implementations, the disclosed systems utilize the segmentation refinement neural network to redetermine whether a set of uncertain pixels corresponds to one or more objects depicted in the digital visual media item based on low-level (e.g., local) feature values extracted from feature maps generated for the digital visual media item.

Abstract: Embodiments are disclosed for generating lens blur effects. The disclosed systems and methods comprise receiving a request to apply a lens blur effect to an image, the request identifying an input image and a first disparity map, generating a plurality of disparity maps and a plurality of distance maps based on the first disparity map, splatting influences of pixels of the input image using a plurality of reshaped kernel gradients, gathering aggregations of the splatted influences, and determining a lens blur for a first pixel of the input image in an output image based on the gathered aggregations of the splatted influences.

Abstract: The present disclosure relates to systems, methods, and non-transitory computer-readable media that generate refined segmentation masks for digital visual media items. For example, in one or more embodiments, the disclosed systems utilize a segmentation refinement neural network to generate an initial segmentation mask for a digital visual media item. The disclosed systems further utilize the segmentation refinement neural network to generate one or more refined segmentation masks based on uncertainly classified pixels identified from the initial segmentation mask. To illustrate, in some implementations, the disclosed systems utilize the segmentation refinement neural network to redetermine whether a set of uncertain pixels corresponds to one or more objects depicted in the digital visual media item based on low-level (e.g., local) feature values extracted from feature maps generated for the digital visual media item.

Abstract: The present disclosure describes systems, non-transitory computer-readable media, and methods for utilizing a machine learning model trained to determine subtle pose differentiations to analyze a repository of captured digital images of a particular user to automatically capture digital images portraying the user. For example, the disclosed systems can utilize a convolutional neural network to determine a pose/facial expression similarity metric between a sample digital image from a camera viewfinder stream of a client device and one or more previously captured digital images portraying the user. The disclosed systems can determine that the similarity metric satisfies a similarity threshold, and automatically capture a digital image utilizing a camera device of the client device. Thus, the disclosed systems can automatically and efficiently capture digital images, such as selfies, that accurately match previous digital images portraying a variety of unique facial expressions specific to individual users.

Abstract: The present disclosure relates to systems, methods, and non-transitory computer-readable media that generate refined segmentation masks for digital visual media items. For example, in one or more embodiments, the disclosed systems utilize a segmentation refinement neural network to generate an initial segmentation mask for a digital visual media item. The disclosed systems further utilize the segmentation refinement neural network to generate one or more refined segmentation masks based on uncertainly classified pixels identified from the initial segmentation mask. To illustrate, in some implementations, the disclosed systems utilize the segmentation refinement neural network to redetermine whether a set of uncertain pixels corresponds to one or more objects depicted in the digital visual media item based on low-level (e.g., local) feature values extracted from feature maps generated for the digital visual media item.

Abstract: Facial skin mask generated by a digital content creation system is described. The digital content creation system includes digital effects on skin in facial regions of digital content with efficiency and accuracy. Upon identifying a facial region within digital content, the system generates a first regional skin mask, a second regional skin mask, and combines both of the first and second regional skin masks to generate a facial skin mask indicative of skin of the identified facial regions depicted in digital content. The digital content creation system then modifies digital content by applying user selected digital effects to the skin of the facial region using the generated facial skin mask.

Abstract: Facial skin mask generated by a digital content creation system is described. The digital content creation system includes digital effects on skin in facial regions of digital content with efficiency and accuracy. Upon identifying a facial region within digital content, the system generates a first regional skin mask, a second regional skin mask, and combines both of the first and second regional skin masks to generate a facial skin mask indicative of skin of the identified facial regions depicted in digital content. The digital content creation system then modifies digital content by applying user selected digital effects to the skin of the facial region using the generated facial skin mask.

Abstract: This application relates generally to parallel computer processing, and more specifically, to parallel processing within a rendering engine via parallel scene graphs. One or more parallel scene graphs or parallel data graphs may be provided to a rendering engine. The rendering engine may identify dependencies within the parallel data structures and process, in parallel, one or more aspects of a the parallel data structure.

Abstract: This application relates generally to parallel computer processing, and more specifically, to parallel processing within a rendering engine via parallel scene graphs. One or more parallel scene graphs or parallel data graphs may be provided to a rendering engine. The rendering engine may identify dependencies within the parallel data structures and process, in parallel, one or more aspects of a the parallel data structure.