-
公开(公告)号：US11688100B1
公开(公告)日：2023-06-27
申请号：US17351869
申请日：2021-06-18
Applicant: Apple Inc.
Inventor： Libin Sun , Feng Li , Jianping Zhou , Xiaoxing Li , Jia Xue
Abstract: Devices, methods, and non-transitory program storage devices are disclosed to provide enhanced images in multi-camera systems, e.g., by using information from images captured by cameras with different properties in terms of optics and/or sensors. In one embodiment, the techniques comprise: obtaining a first image from a first image capture device, wherein the first image has a first field of view (FOV) and a first set of quality characteristics; obtaining a second image from a second image capture device, wherein the second image has a second FOV and a second set of quality characteristics, and wherein the second FOV partially overlaps the first FOV; obtaining a neural network that produces a modified second image having a modified second set of quality characteristics determined by the neural network attempting to match the first set of quality characteristics; and generating an output image based, at least in part, on the modified second image.
-
公开(公告)号：US12205249B2
公开(公告)日：2025-01-21
申请号：US17805071
申请日：2022-06-02
Applicant: Apple Inc.
Inventor： Libin Sun
Abstract: Devices, methods, and non-transitory program storage devices are disclosed to provide enhanced synthetic Shallow Depth of Field (SDOF) images, e.g., by using information from images captured by image camera devices and one or more Deep Neural Networks (DNNs) trained to determine how much blurring should be applied to pixels in a mask region (e.g., a region of pixels having an indeterminate foreground or background status and threshold level of gradient magnitude) within an image, given context from surrounding pixels in a reference image and/or an unenhanced synthetic SDOF image. To train the DNNs, various sets of ground truth DSLR images of static scenes, captured at varying aperture settings, may be analyzed. Preferably, such static scenes cover many examples of human subjects (or other objects of interest) with different amounts of scene foreground/background separation, lighting conditions, and various kinds of hair, fabric, or other fine-grained details occurring near their foreground/background transition.
-
公开(公告)号：US20220398704A1
公开(公告)日：2022-12-15
申请号：US17805071
申请日：2022-06-02
Applicant: Apple Inc.
Inventor： Libin Sun
Abstract: Devices, methods, and non-transitory program storage devices are disclosed to provide enhanced synthetic Shallow Depth of Field (SDOF) images, e.g., by using information from images captured by image camera devices and one or more Deep Neural Networks (DNNs) trained to determine how much blurring should be applied to pixels in a mask region (e.g., a region of pixels having an indeterminate foreground or background status and threshold level of gradient magnitude) within an image, given context from surrounding pixels in a reference image and/or an unenhanced synthetic SDOF image. To train the DNNs, various sets of ground truth DSLR images of static scenes, captured at varying aperture settings, may be analyzed. Preferably, such static scenes cover many examples of human subjects (or other objects of interest) with different amounts of scene foreground/background separation, lighting conditions, and various kinds of hair, fabric, or other fine-grained details occurring near their foreground/background transition.
-
-
Search Results
3
records
(took 0.023 seconds)
