Abstract: Systems and techniques are described herein for processing video data. In some examples, a process is described that can include obtaining a plurality of frames, determining a scene cut in the plurality of frames, and determining a smoothed histogram based on the determined scene cut. For instance, the process can include determining a first characteristic of at least a first frame of the plurality of frames and a second characteristic of at least a second frame of the plurality of frames, determining whether a difference between the first characteristic and the second characteristic is greater than a threshold difference, and determining the scene cut based a determination that the difference between the first characteristic and the second characteristic is greater than the threshold difference.

Abstract: Image frames for computational photography may be corrected, such as through rolling shutter correction (RSC), prior to fusion of the image frames to reduce wobble and jitter artifacts present in a video sequence of HDR-enhanced image frames. First and second motion data regarding motion of the image capture device may be determined for times corresponding to the capturing of the first and second image frames, respectively. The rolling shutter correction (RSC) may be applied to the first and second image frames based on both the first and second motion data. The corrected first and second image frames may then be aligned and fused to obtain a single output image frame with higher dynamic range than either of the first or second image frames.

Abstract: Apparatus for generating images using a mechanical infrared cut-off switch are disclosed herein. An example apparatus including an optical system for generating images includes a mechanical infrared light filter movable between a first position and a second position. The mechanical infrared light filter may be configured to allow infrared light to pass through the optical system while in the first position and configured to filter out infrared light from the optical system while in the second position. The example apparatus also includes an imaging sensor including imaging pixels and infrared pixels, and may be configured to receive light from the optical system. Additionally, the example apparatus includes a processor configured to receive visible light data and infrared light data from the image sensor. The processor may further be configured to generate a combined image based on the visible light data and the infrared light data.

Abstract: A transformer may transform image data from a first color pattern to a second color pattern. The transforming of image data may be applied to image data received from a memory storing an array of intensities corresponding to a first color pattern of a first color filter array (CFA) of an image sensor to a second color pattern. The second color pattern may be a color pattern of a size smaller that the first CFA. Remosaicing may be applied to the second color pattern to obtain image data organized in a Bayer color pattern. The transforming may be configured to operate on data from an image sensor to obtain different zoom levels not available without applying a digital zoom algorithm that involve upscaling, which reduces the image quality of the image data.

Abstract: Systems and techniques are described for generating a high dynamic range (HDR) image. An imaging system can be configured to receive a first image captured by an image sensor according to a first exposure time. The imaging system can generate a modified image based on the first image by modifying the first image using a gain setting to simulate a second exposure time based on exposure compensation. The imaging system generates a high dynamic range (HDR) image at least in part by merging multiple images. The multiple images include a second-exposure image that corresponds to the second exposure time. The second-exposure image can be the modified image. The second-exposure image can be based on the modified image, processed variant of the modified image processed for noise reduction based on one or more other images actually captured using the second exposure time.

Abstract: Image frames for computational photography may be corrected, such as through rolling shutter correction (RSC), prior to fusion of the image frames to reduce wobble and jitter artifacts present in a video sequence of HDR-enhanced image frames. First and second motion data regarding motion of the image capture device may be determined for times corresponding to the capturing of the first and second image frames, respectively. The rolling shutter correction (RSC) may be applied to the first and second image frames based on both the first and second motion data. The corrected first and second image frames may then be aligned and fused to obtain a single output image frame with higher dynamic range than either of the first or second image frames.

Abstract: The present disclosure relates to methods and apparatus for image processing. For example, disclosed techniques facilitate a sparse infrared pixel design for image sensors. Aspects of this disclosure include an image sensor including a pattern of pixels, the pattern of pixels including a set of visible light pixels and a set of infrared light pixels. The image sensor may be configured to generate visible light data and infrared light data based on received light. Aspects of the present disclosure also include a processor coupled to the image sensor. The processor may be configured to generate a configurable mask based on the pattern of pixels of the image sensor, and where values of the configurable mask correspond to pixel locations of the image sensor.

Abstract: Methods, systems, and devices for anchor frame switching in an imaging system are described. A device may generate a pixel map based on a set of frames. A first subset of frames of the set of frames have a first exposure and a second subset of frames of the set of frames have a second exposure different than the first exposure. The device may determine a region of the pixel map representing motion between at least two frames of the first set of frames. If the device determines that a quantity of motion pixels in the region is higher than a threshold, the device may select a short exposure frame as an anchor frame, beginning at a later frame. Otherwise, the device may maintain using a long exposure frame as the anchor frame.

Abstract: Methods, systems, and devices for minimizing ghosting in high dynamic range image processing are described. The method may include capturing from a sensor of the device a downscaled first frame of a first exposure length and a downscaled second frame of a second exposure length that is longer than the first exposure length, identifying a highlight region associated with the downscaled first frame and a motion region associated with the downscaled first frame and with the downscaled second frame, blending the motion region in accordance with determining whether at least a portion of the motion region overlaps the highlight region, and incorporating the blending of the motion region in a set of full resolution frames.

Abstract: Systems, methods, and computer-readable media are provided for foreground image segmentation. In some examples, a method can include obtaining a first image of a target and a second image of the target, the first image having a first field-of-view (FOV) and the second image having a second FOV; determining, based on the first image, a first segmentation map that identifies a first estimated foreground region in the first image; determining, based on the second image, a second segmentation map that identifies a second estimated foreground region in the second image; generating a third segmentation map based on the first segmentation map and the second segmentation map; and generating, using the second segmentation map and the third segmentation map, a refined segmentation mask that identifies the target as a foreground region of the first and/or second image.

Abstract: Systems, methods, and computer-readable media are provided for foreground image segmentation. In some examples, a method can include obtaining a first image of a target and a second image of the target, the first image having a first field-of-view (FOV) and the second image having a second FOV; determining, based on the first image, a first segmentation map that identifies a first estimated foreground region in the first image; determining, based on the second image, a second segmentation map that identifies a second estimated foreground region in the second image; generating a third segmentation map based on the first segmentation map and the second segmentation map; and generating, using the second segmentation map and the third segmentation map, a refined segmentation mask that identifies the target as a foreground region of the first and/or second image.

Abstract: Aspects of the present disclosure relate to QCFA deblurring. An example method includes obtaining an image to be processed and determining an average of pixel values between a first one or more pixels and a second one or more pixels of the image. The second one or more pixels neighbor the first one or more pixels. The method also includes determining a difference between pixel values of the first one or more pixels and the second one or more pixels, generating one or more weights based on the average and the difference, and combining the average and the difference based on the one or more weights to generate a deblurred pixel value. A processed image includes one or more deblurred pixel values.

Abstract: The present disclosure relates to methods and apparatus for image processing. For example, disclosed techniques facilitate a sparse infrared pixel design for image sensors. Aspects of this disclosure include an image sensor including a pattern of pixels, the pattern of pixels including a set of visible light pixels and a set of infrared light pixels. The image sensor may be configured to generate visible light data and infrared light data based on received light. Aspects of the present disclosure also include a processor coupled to the image sensor. The processor may be configured to generate a configurable mask based on the pattern of pixels of the image sensor, and where values of the configurable mask correspond to pixel locations of the image sensor.

Abstract: Apparatus for generating images using a mechanical infrared cut-off switch are disclosed herein. An example apparatus including an optical system for generating images includes a mechanical infrared light filter movable between a first position and a second position. The mechanical infrared light filter may be configured to allow infrared light to pass through the optical system while in the first position and configured to filter out infrared light from the optical system while in the second position. The example apparatus also includes an imaging sensor including imaging pixels and infrared pixels, and may be configured to receive light from the optical system. Additionally, the example apparatus includes a processor configured to receive visible light data and infrared light data from the image sensor. The processor may further be configured to generate a combined image based on the visible light data and the infrared light data.

Abstract: Systems, methods, and computer-readable media are provided for generating an image processing effect via disparity-guided salient object detection. In some examples, a system can detect a set of superpixels in an image; identify, based on a disparity map generated for the image, an image region containing at least a portion of a foreground of the image; calculate foreground queries identifying superpixels in the image region having higher saliency values than other superpixels in the image region; rank a relevance between each superpixel and one or more foreground queries; generate a saliency map for the image based on the ranking of the relevance between each superpixel and the one or more foreground queries; and generate, based on the saliency map, an output image having an effect applied to a portion of the output image.

Abstract: Aspects of the present disclosure relate to systems and methods for selectively processing tiles of a captured image. The captured image may be parsed into a plurality of tiles. A subset of the plurality of tiles may be selected based at least in part on at least one metric. The selected tiles may be processed according to a first technique, while each unselected tile may be selectively processed according to one or more techniques different from the first technique. The first technique and the one or more techniques may each be either an image demosaicing technique or an image denoising technique.

Abstract: Methods, systems, and devices for processing color information are described. The method may include receiving a first set of image data collected at an image sensor based on a first color filter array having a first filter pattern, calculating a color saturation gradient based on the first set of image data, and calculating multiple color gradients based on the color saturation gradient. The multiple color gradients may each be associated with a region of the first set of image data, where each color gradient characterizes a color variation along a different direction of the region. The method may include generating a second set of image data including a color value for the region based on comparing a first direction gradient of the multiple color gradients with a second direction gradient of the multiple color gradients. Color values associated with the regions may be mapped to a second color filter array.

Abstract: In general, techniques are described regarding a combined monochrome and chromatic camera sensor. A camera comprising a camera sensor and a processor may be configured to perform the techniques. The camera sensor may include pixel sensors, monochrome filters disposed over a first subset of the pixel sensors, and color filters disposed over a second subset of the pixel sensors. The first subset of the pixel sensors may be configured to capture a monochrome image of a scene. The second subset of the pixel sensors may be configured to capture, concurrently with the capture of the monochrome image, a color image of the scene, where a number of the first subset of the pixel sensors is greater than a number of the second subset of the pixel sensors. The processor may be configured to process the monochrome image and the color image to obtain an enhanced color image of the scene.

Abstract: Methods, systems, and devices for image processing are described. A device may capture image data based on a color filter array (CFA) associated with an image sensor. The image data may include a set of pixels. The device may determine a CFA pattern of the CFA and select a configuration appropriately. The configuration may include an indication of a first set of neighboring pixels for each pixel of the set of pixels to use to determine that a pixel is defective and a second set of neighboring pixels for each pixel of the set of pixels to use to correct the defective pixel. The device may determine that the pixel of the set of pixels is defective using the configuration, and correct the defective pixel using pixel values of the first set of neighboring pixels, or pixel values of the second set of neighboring pixels, or both.

Abstract: Methods, systems, and devices for processing color information are described. The method may include receiving a first set of image data collected at an image sensor based on a first color filter array having a first filter pattern, calculating a color saturation gradient based on the first set of image data, and calculating multiple color gradients based on the color saturation gradient. The multiple color gradients may each be associated with a region of the first set of image data, where each color gradient characterizes a color variation along a different direction of the region. The method may include generating a second set of image data including a color value for the region based on comparing a first direction gradient of the multiple color gradients with a second direction gradient of the multiple color gradients. Color values associated with the regions may be mapped to a second color filter array.