Abstract: Visual content that includes spatial portions is obtained. A determination is made that one of the spatial portions includes a face. Based on the determination, encoding quality parameters for the one of the spatial portions is identified. The encoding quality parameters are obtained by combining a first distortion model related to the obtaining the visual content with a second model that emphasizes the one of the spatial portions The visual content is encoded. The encoding quality parameters are stored, in association with but separate from, the one of the spatial portions. After decoding, the one of the spatial portions are rendered based on the encoding quality parameters. The encoding quality parameters are obtained by combining a first distortion model related to the obtaining the visual content with a second model that emphasizes the one of the spatial portions.

Abstract: Encoded content is accessed. The encoded content includes an encoded first centrally located tile corresponding to a first centrally located tile of a first image, an encoded first peripherally located tile of the first image, and an encoded second peripherally located tile of a second image. The encoded first peripherally located tile is decoded to obtain a decoded first peripherally located tile. The encoded second peripherally located tile is decoded to obtain a decoded second peripherally located tile. The decoded first peripherally located tile and the decoded second peripherally located tile are stitched to obtain a stitched image portion. The stitched image portion is encoded to obtain an encoded stitched image portion. An encoded stitched image of the first image and the second image is obtained by combining the encoded first centrally located tile, and the encoded stitched image portion.

Abstract: A method includes obtaining visual content comprising spatial portions; determining respective spatial qualities of the spatial portions, wherein the respective spatial qualities are based on locations of the spatial portions within the visual content; and encoding the spatial portions of the visual content based on the respective spatial qualities. An apparatus includes a camera, a display, and a processor. The processor is configured to identify, using facial recognition, a face of a user of the apparatus; identify a distance of the face of the user to the display; and render visual content on the display using a quality that is based on the distance.

Abstract: A method for encoding images includes decoding a first encoded image to obtain a first decoded image, where the first decoded image includes a first decoded portion corresponding to a first encoded portion of the first encoded image and a second decoded portion corresponding to a second encoded portion of the first encoded image; decoding a second encoded image to obtain a second decoded image; combining the first decoded image and the second decoded image to obtain a single decoded image; and encoding the single decoded image to obtain a single encoded image that includes a third and a fourth encoded portions. Encoding the single decoded image includes obtaining the third encoded portion of the single encoded image by copying the first encoded portion of the first encoded image; and obtaining the fourth encoded portion of the single encoded image by encoding the second decoded portion using an encoder.

Abstract: A processing device generates composite images from a sequence of images. The composite images may be used as frames of video. A foreground/background segmentation is performed at selected frames to extract a plurality of foreground object images depicting a foreground object at different locations as it moves across a scene. The foreground object images are stored to a foreground object list. The foreground object images in the foreground object list are overlaid onto subsequent video frames that follow the respective frames from which they were extracted, thereby generating a composite video.

Abstract: Images captured by multi-camera arrays with overlap regions can be stitched together using image stitching operations. An image stitching operation can be selected for use in stitching images based on a number of factors. An image stitching operation can be selected based on a view window location of a user viewing the images to be stitched together. An image stitching operation can also be selected based on a type, priority, or depth of image features located within an overlap region. Finally, an image stitching operation can be selected based on a likelihood that a particular image stitching operation will produce visible artifacts. Once a stitching operation is selected, the images corresponding to the overlap region can be stitched using the stitching operation, and the stitched image can be stored for subsequent access.

Abstract: A system and methods are disclosed for optimal format selection for video players based on visual quality. The method includes generating a plurality of reference transcoded versions of a reference video, obtaining quality scores for frames of the plurality of reference transcoded versions of the reference video, generating a first training input comprising a set of color attributes, spatial attributes, and temporal attributes of the frames of the reference video, and generating a first target output for the first training input, wherein the first target output comprises the quality scores for the frames of the plurality of reference transcoded versions of the reference video. The method further includes providing the training data to train a machine learning model on (i) a set of training inputs comprising the first training input and (ii) a set of target outputs comprising the first target output.

Abstract: A target image captured from a fisheye lens or other lens with known distortion parameters may be transformed to align it to a reference image. Corresponding features may be detected in the target image and the reference image. The features may be transformed to a spherical coordinate space. In the spherical space, images may be re-pointed or rotated in three dimensions to align all or a subset of the features of the target image to the corresponding features of the reference image. For example, in a sequence of images, background features of the target image in the spherical image space may be aligned to background features of the reference image in the spherical image space to compensate for camera motion while preserving foreground motion. An inverse transformation may then be applied to bring the images back into the original image space.

Abstract: Systems and methods for providing panoramic image and/or video content using multi-resolution stitching. Panoramic content may include stitched spherical (360-degree) images and/or VR video. In some implementations, multi-resolution stitching functionality may be embodied in a spherical image capture device that may include two lenses configured to capture pairs of hemispherical images. The capture device may obtain images (e.g., representing left and right hemispheres) that may be characterized by 180-degree (or greater) field of view. Source images may be combined using multi-resolution stitching methodology. Source images may be transformed to obtain multiple image components characterized by two or more image resolutions.

Abstract: Systems and methods for modifying image distortion (curvature) for viewing distance in post capture. Presentation of imaging content on a content display device may be characterized by a presentation field of view (FOV). Presentation FOV may be configured based on screen dimensions of the display device and distance between the viewer and the screen. Imaging content may be obtained by an activity capture device characterized by a wide capture field of view lens (e.g., fish-eye). Images may be transformed into rectilinear representation for viewing. When viewing images using presentation FOV that may narrower than capture FOV, transformed rectilinear images may appear distorted. A transformation operation may be configured to account for presentation FOV—capture FOV mismatch. In some implementations, the transformation may include fish-eye to rectilinear transformation characterized by a transformation strength that may be configured based on a ratio of the presentation FOV to the capture FOV.

Abstract: A device includes a processor that is configured to obtain first facets of a first wide field-of-view image. An object is identified in a facet of the first facets. A second wide field-of-view image is obtained. A location of the object is identified in the second wide field-of-view image. Using the location of the object, the second wide field-of-view image is partitioned into second facets such that no boundary of any of the second facets overlaps the object. The second facets are then encoded in a compressed bitstream.

Abstract: Apparatus and methods for characterizing panoramic content, such as by a wide field of view and large image size. In one implementation, a panoramic image may be mapped to a cube or any other projection e.g icosahedron or octahedron. The disclosure exploits content continuity between facets, such as in the case of encoding/decoding cube-projected images. One facet may be encoded/decoded independently from other facets to obtain a seed facet. One or more transformed versions of the seed facet may be obtained; e.g., one corresponding to a 90° counterclockwise rotation, another to a 90° clockwise rotation, and one to an 180° rotation. Transformed versions may be used to form an augmented image. The remaining facets of the cube may be encoded using transformed versions within the augmented image.

Abstract: A processing device generates composite images from a sequence of images. The composite images may be used as frames of video. A foreground/background segmentation is performed at selected frames to extract a plurality of foreground object images depicting a foreground object at different locations as it moves across a scene. The foreground object images are stored to a foreground object list. The foreground object images in the foreground object list are overlaid onto subsequent video frames that follow the respective frames from which they were extracted, thereby generating a composite video.

Abstract: A method for encoding images includes decoding a first encoded image to obtain a first decoded image, where the first decoded image includes a first decoded portion corresponding to a first encoded portion of the first encoded image and a second decoded portion corresponding to a second encoded portion of the first encoded image; decoding a second encoded image to obtain a second decoded image; combining the first decoded image and the second decoded image to obtain a single decoded image; and encoding the single decoded image to obtain a single encoded image that includes a third and a fourth encoded portions. Encoding the single decoded image includes obtaining the third encoded portion of the single encoded image by copying the first encoded portion of the first encoded image; and obtaining the fourth encoded portion of the single encoded image by encoding the second decoded portion using an encoder.

Abstract: Images captured by multi-camera arrays with overlap regions can be stitched together using image stitching operations. An image stitching operation can be selected for use in stitching images based on a number of factors. An image stitching operation can be selected based on a view window location of a user viewing the images to be stitched together. An image stitching operation can also be selected based on a type, priority, or depth of image features located within an overlap region. Finally, an image stitching operation can be selected based on a likelihood that a particular image stitching operation will produce visible artifacts. Once a stitching operation is selected, the images corresponding to the overlap region can be stitched using the stitching operation, and the stitched image can be stored for subsequent access.

Abstract: Systems and methods for providing panoramic image and/or video content using spatially selective encoding and/or decoding. Panoramic content may include stitched spherical (360-degree) images and/or VR video. In some implementations, selective encoding functionality may be embodied in a spherical image capture device that may include two lenses configured to capture pairs of hemispherical images. Encoded source images may be decoded and stitched in order to obtain a combined image characterized by a greater field of view as compared to source images. The stitched image may be encoded using a selective encoding methodology including: partitioning a stitched image into multiple portions, determining if a portion is to be re-encoded. If the image portion is to be re-encoded, re-encoding the image portion. If an image portion is not to be re-encoded, copying previously encoded image portion in lieu of encoding.

Abstract: Systems and methods for providing panoramic image and/or video content using multi-resolution stitching. Panoramic content may include stitched spherical (360-degree) images and/or VR video. In some implementations, multi-resolution stitching functionality may be embodied in a spherical image capture device that may include two lenses configured to capture pairs of hemispherical images. The capture device may obtain images (e.g., representing left and right hemispheres) that may be characterized by 180-degree (or greater) field of view. Source images may be combined using multi-resolution stitching methodology. Source images may be transformed to obtain multiple image components characterized by two or more image resolutions.

Abstract: A processing device generates composite images from a sequence of images. The composite images may be used as frames of video. A foreground/background segmentation is performed at selected frames to extract a plurality of foreground object images depicting a foreground object at different locations as it moves across a scene. The foreground object images are stored to a foreground object list. The foreground object images in the foreground object list are overlaid onto subsequent video frames that follow the respective frames from which they were extracted, thereby generating a composite video.

Abstract: A method includes obtaining visual content comprising spatial portions; determining respective spatial qualities of the spatial portions, wherein the respective spatial qualities are based on locations of the spatial portions within the visual content; and encoding the spatial portions of the visual content based on the respective spatial qualities. An apparatus includes a camera, a display, and a processor. The processor is configured to identify, using facial recognition, a face of a user of the apparatus; identify a distance of the face of the user to the display; and render visual content on the display using a quality that is based on the distance.

Abstract: Images captured by multi-camera arrays with overlap regions can be stitched together using image stitching operations. An image stitching operation can be selected for use in stitching images based on a number of factors. An image stitching operation can be selected based on a view window location of a user viewing the images to be stitched together. An image stitching operation can also be selected based on a type, priority, or depth of image features located within an overlap region. Finally, an image stitching operation can be selected based on a likelihood that a particular image stitching operation will produce visible artifacts. Once a stitching operation is selected, the images corresponding to the overlap region can be stitched using the stitching operation, and the stitched image can be stored for subsequent access.