Abstract: A method includes obtaining multiple image frames of a scene using at least one camera of an electronic device. The method also includes using a convolutional neural network to generate blending maps associated with the image frames. The blending maps contain or are based on both a measure of motion in the image frames and a measure of how well exposed different portions of the image frames are. The method further includes generating a final image of the scene using at least some of the image frames and at least some of the blending maps. The final image of the scene may be generated by blending the at least some of the image frames using the at least some of the blending maps, and the final image of the scene may include image details that are lost in at least one of the image frames due to over-exposure or under-exposure.

Abstract: A method of processing image data in a composite image is described. The method comprises analyzing a multi-level blending of images of the composite image; identifying sharp transition boundary areas of the composite image; applying less attenuation for higher levels of multi-level blending of the composite image in the sharp transition boundary areas; and applying more attenuation for lower levels of multi-level blending of the composite image in the sharp transition boundary areas.

Abstract: A method includes obtaining multiple image frames of a scene using at least one camera of an electronic device and processing the multiple image frames to generate a higher-resolution image of the scene. Processing the multiple image frames includes generating an initial estimate of the scene based on the multiple image frames. Processing the multiple image frames also includes, in each of multiple iterations, (i) generating a current estimate of the scene based on the image frames and a prior estimate of the scene and (ii) regularizing the generated current estimate of the scene. The regularized current estimate of the scene from one iteration represents the prior estimate of the scene in a subsequent iteration. The iterations continue until the estimates of the scene converge on the higher-resolution image of the scene.

Abstract: A method includes capturing multiple ambient images of a scene using at least one camera of an electronic device and without using a flash of the electronic device. The method also includes capturing multiple flash images of the scene using the at least one camera of the electronic device and during firing of a pilot flash sequence using the flash. The method further includes analyzing multiple pairs of images to estimate exposure differences obtained using the flash, where each pair of images includes one of the ambient images and one of the flash images that are both captured using a common camera exposure and where different pairs of images are captured using different camera exposures. In addition, the method includes determining a flash strength for the scene based on the estimate of the exposure differences and firing the flash based on the determined flash strength.

Abstract: A method includes capturing multiple pairs of images of a scene at different exposures using at least one camera of an electronic device. Each pair of images includes (i) an ambient image of the scene captured without using a flash of the electronic device and (ii) a flash image of the scene captured using the flash of the electronic device. The method also includes rendering a final image of the scene with a bokeh that is determined using the multiple pairs of images. One of the ambient images or the flash images are captured in order of increasing exposure time, and the other of the ambient images or the flash images are captured in order of decreasing exposure time. The method may also include estimating a depth map associated with the scene using the pairs of images, where the bokeh is based on the depth map.

Abstract: A method of generating an image from multiple cameras having different focal lengths is described. The method comprising receiving a wide image and a tele image; aligning the wide image and the tele image to overlap a common field of view; correcting for photometric differences between the wide image and the tele image; selecting a stitching seam for the wide image and the tele image; and joining the wide image and the tele image to generate a composite image, wherein a first portion of the composite image on one side of the stitching seam is from the wide image and a second portion of the composite image on the other side of the stitching seam is from the tele image. An electronic device for generating an image is also described.

Abstract: A method to enhance quality of an image comprises receiving the image; identifying a region of the image associated with an object in the image; applying first image processing for the object in the image; applying second image processing on the remaining portions of the image.

Abstract: A method to enhance quality of an image is described. The method comprises receiving the image; identifying a region of the image having skin; performing motion analysis in the region of the image having skin; and if motion is detected, then controlling blending in the region of the image having skin to avoid blurring of texture in the skin.

Abstract: A method to enhance quality of an image is described. The method comprises receiving the image; identifying a region of the image having skin; performing motion analysis in the region of the image having skin; and if motion is detected, then controlling blending in the region of the image having skin to avoid blurring of texture in the skin.

Abstract: Traditionally, providing parallel processing within a multi-core system has been very difficult. Here, however, a system is provided where serial source code is automatically converted into parallel source code, and a processing cluster is reconfigured “on the fly” to accommodate the parallelized code based on an allocation of memory and compute resources. Thus, the processing cluster and its corresponding system programming tool provide a system that can perform parallel processing from a serial program that is transparent to a user. Generally, a control node connected to the address and data leads of a host processor uses messages to control the processing of data in a processing cluster. The cluster includes nodes of parallel processors, shared function memory, a global load/store, and hardware accelerators all connected to the control node by message busses. A crossbar data interconnect routes data to the cluster circuits separate from the message busses.

Abstract: A method includes identifying feature points in an image in images generated of a scene by a camera and identifying locations of the identified feature points in remaining images in the images. The method also includes selecting a group of the identified feature points indicative of relative motion of the camera between image captures and aligning a set of the images using the selected group of feature points. The method may further include selecting a reference image from the set of aligned images, weighting other images the set, and combining the reference image with the weighted images. Weighting of the other images may include, for each other image in the set, comparing the other image and the reference image to identify one or more moving objects in the other image and applying a weight to pixel locations in the other image.

Abstract: A mobile device includes an embedded digital camera that is configured to capture a burst of N images. The mobile device includes processing circuitry comprising a registration module configured to, for each image within the burst of images: analyze an amount of warp of the image and generate a set of affine matrices indicating the amount of warp of the image. The processing circuitry includes a High Fidelity Interpolation block configured to, for each image within the burst of images: perform affine transformation using the set of affine matrices associated with the image, apply an aliasing-retaining interpolation filter, and implement rotation transformation and sub-pixel shifts, yielding an interpolated image. The processing circuitry includes a blending module configured to receive the N interpolated images and blend the N interpolated images into a single-blended image having a user-selected digital zoom ratio.

Abstract: A mobile device includes an embedded digital camera that is configured to capture a burst of N images. The mobile device includes processing circuitry comprising a registration module configured to, for each image within the burst of images: analyze an amount of warp of the image and generate a set of affine matrices indicating the amount of warp of the image. The processing circuitry includes a High Fidelity Interpolation block configured to, for each image within the burst of images: perform affine transformation using the set of affine matrices associated with the image, apply an aliasing-retaining interpolation filter, and implement rotation transformation and sub-pixel shifts, yielding an interpolated image. The processing circuitry includes a blending module configured to receive the N interpolated images and blend the N interpolated images into a single-blended image having a user-selected digital zoom ratio.

Abstract: A method and apparatus for aligning and combining images. The method includes identifying feature points in an image in images generated of a scene by a camera and identifying locations of the identified feature points in remaining images in the images. The method also includes selecting a group of the identified feature points indicative of relative motion of the camera between image captures and aligning a set of the images using the selected group of feature points. The method may further include selecting a reference image from the set of aligned images, weighting other images the set, and combining the reference image with the weighted images. Weighting of the other images may include, for each other image in the set, comparing the other image and the reference image to identify one or more moving objects in the other image and applying a weight to pixel locations in the other image.

Abstract: Traditionally, providing parallel processing within a multi-core system has been very difficult. Here, however, a system in provided where serial source code is automatically converted into parallel source code, and a processing cluster is reconfigured “on the fly” to accommodate the parallelized code based on an allocation of memory and compute resources. Thus, the processing cluster and its corresponding system programming tool provide a system that can perform parallel processing from a serial program that is transparent to a user.

Abstract: Digital Still Camera (DSC) includes separate preview engine, burst mode compression/decompression engine, image pipeline, CCD plus CCD controller, and memory plus memory controller. ARM microprocessor and DSP share control. Color filter array interpolation by use of green high-frequency added to red and blue plus interpolation.

Abstract: Digital Still Camera (DSC) includes separate preview engine, burst mode compression/decompression engine, image pipeline, CCD plus CCD controller, and memory plus memory controller. ARM microprocessor and DSP share control. Color filter array interpolation by use of green high-frequency added to red and blue plus interpolation.