Abstract: In one embodiment, a computing system may set data to a first group of registers. The first group of registers may be configured to be accessed during a single operation cycle. The system may set a number of patterns to a second group of registers. Each pattern of the number of patterns may include an array of index for the data stored in the first group of registers. The system may select, for a first vector register associated with a vector engine, a first pattern from the patterns stored in the second group of registers. The system may load a first portion of the data from the first group of registers to the first vector register based on the first pattern selected for the first vector register from the patterns stored in the second group of registers.

Abstract: In one embodiment, a system comprising a processor and a non-transitory memory coupled to the processor comprising instructions executable by the processor. The processor, comprising an internal memory; a Multiply-Accumulate (MAC) array; a first vector register array; a second vector register array; and a third vector register array, is operable when executing instructions to transfer weights for M filters and an input activation tensor from an external memory to the internal memory, insert paddings to the input activation tensor in the internal memory based on first configuration parameters, configure the MAC array to a required shape based on second configuration parameters for convolution operations between the input activation tensor and the M filters, and calculate a row of the output activation tensor by performing the convolution operations on corresponding R rows of the input activation tensor with the M filters, wherein R is a filter height.

Abstract: In one embodiment, a system comprising a processor and a non-transitory memory coupled to the processor comprising instructions executable by the processor. The processor, comprising an internal memory; a Multiply-Accumulate (MAC) array; a first vector register array; a second vector register array; and a third vector register array, is operable when executing a first instruction among the instructions to feed a weight vector array from the second vector register array to the MAC array, broadcast an input activation vector to the MAC array, multiply an input activation value broadcast to the MAC unit from the input activation vector and a weight value fed to the MAC unit from the weight vector array at each MAC unit in the MAC array, and store a partial output activation vector to the third vector register array, wherein the partial output activation vector is the output of the MAC array.

Abstract: A method for generating an optical flow for a plurality of successive image frames includes executing an initialization process by performing a plurality of raster scans of a patch of pixels in one or more of the plurality of successive image frames in parallel. The plurality of raster scans of the patch of pixels includes a plurality of optical flow estimates between the plurality of successive image frames. The method includes executing a propagation process based on the plurality of optical flow estimates between the plurality of successive image frames. Executing the propagation process includes propagating the plurality of optical flow estimates for one or more neighboring pixels associated with the patch of pixels. The method includes executing a search process by identifying one or more offsets based on the plurality of optical flow estimates for the one or more neighboring pixels associated with the patch of pixels.

Abstract: An electronic device includes one or more processors and memory storing instructions for execution by the one or more processors. The stored instructions include instructions for: receiving infrared image information for a three-dimensional area; receiving non-infrared image information for the same three-dimensional area; performing nonlinear intensity adjustment for the received infrared image information; performing nonlinear intensity adjustment for the received non-infrared image information; blending the intensity-adjusted infrared image information and the intensity-adjusted non-infrared image information to obtain a merged image information; and providing the merged image information for determining a depth map. Also disclosed are a corresponding method performed by the electronic device and a computer readable storage medium storing instructions for execution by one or more processors of an electronic device.

Abstract: An electronic device includes one or more processors and memory storing instructions for execution by the one or more processors. The stored instructions include instructions for: receiving infrared image information for a three-dimensional area; receiving non-infrared image information for the same three-dimensional area; performing nonlinear intensity adjustment for the received infrared image information; performing nonlinear intensity adjustment for the received non-infrared image information; blending the intensity-adjusted infrared image information and the intensity-adjusted non-infrared image information to obtain a merged image information; and providing the merged image information for determining a depth map. Also disclosed are a corresponding method performed by the electronic device and a computer readable storage medium storing instructions for execution by one or more processors of an electronic device.

Abstract: In one embodiment, a method includes accessing image-sensor data generated by the image sensor, where the image sensor has a color filter array with a pre-determined number of filter sets, and where each set of filters has a single color, splitting the image-sensor data into a pre-determined number of images, where each image corresponds to a portion of the image-sensor data associated with one of the sets of filters, compressing each of the images using an image compression algorithm, and sending the compressed images to a second computing device, where the second computing device is configured to create an output image based on the compressed images.

Abstract: In one embodiment, a method includes accessing image-sensor data generated by the image sensor, where the image sensor has a color filter array with a pre-determined number of filter sets, and where each set of filters has a single color, splitting the image-sensor data into a pre-determined number of images, where each image corresponds to a portion of the image-sensor data associated with one of the sets of filters, compressing each of the images using an image compression algorithm, and sending the compressed images to a second computing device, where the second computing device is configured to create an output image based on the compressed images.

Abstract: A special-purpose hardware device for mitigating motion-to-photon latency in head-mounted displays may include an image signal processor that receives at least one image frame captured by a camera device of a head-mounted-display system. The special-purpose hardware device may also include an input-formatting component that receives the computer-generated imagery. The special-purpose hardware device may further include a blending component that generates at least one mixed-reality frame by overlaying the computer-generated imagery onto the image frame. The special-purpose hardware device may additionally include a frame-output interface that feeds the mixed-reality frame generated by the blending component to a display device of the head-mounted-display system to facilitate displaying the mixed-reality frame for presentation to a user wearing the head-mounted-display system. Various other apparatuses, systems, and methods are also disclosed.

Abstract: A special-purpose hardware device for achieving intraframe image processing in head-mounted displays may include an image-correction component that receives a single image frame destined to be displayed for presentation to a user wearing a head-mounted-display system. The image-correction component may also receive user-motion data indicating that the user wearing the head-mounted-display system has made at least one movement since generation of the single image frame. The image-correction component may further perform hardware-accelerated intraframe processing operations on the single image frame in accordance with the user-motion data to compensate for the movement made by the user on a line-by-line basis. In addition, the special-purpose hardware device may include a frame-output interface that feeds the single image frame to a display device of the head-mounted-display system to facilitate displaying the single image frame for presentation to the user.

Abstract: The disclosed special-purpose hardware device may include an image signal processor that receives, from a camera device of a head-mounted-display system, image frames of a physical environment. The special-purpose hardware device may also include a positional tracking component that (1) stores at least a portion of the image frames in a cache of the special-purpose hardware device that has a faster access speed than a main memory of the special-purpose hardware device, (2) tracks, based on the portion of the image frames stored in the cache, a change in the position of the head-mounted display system within the physical environment, and (3) stores the change in the position of the head-mounted-display system in the main memory for use in generating one or more augmented-reality frames. The special-purpose hardware device may further include a frame-output interface that feeds the augmented-reality frames to a display device of the head-mounted-display system.

Abstract: A special-purpose hardware device for mitigating motion-to-photon latency in head-mounted displays may include an image signal processor that receives at least one image frame captured by a camera device of a head-mounted-display system. The special-purpose hardware device may also include an input-formatting component that receives the computer-generated imagery. The special-purpose hardware device may further include a blending component that generates at least one mixed-reality frame by overlaying the computer-generated imagery onto the image frame. The special-purpose hardware device may additionally include a frame-output interface that feeds the mixed-reality frame generated by the blending component to a display device of the head-mounted-display system to facilitate displaying the mixed-reality frame for presentation to a user wearing the head-mounted-display system. Various other apparatuses, systems, and methods are also disclosed.

Abstract: A special-purpose hardware device for mitigating motion-to-photon latency in head-mounted displays may include an image signal processor that receives at least one image frame captured by a camera device of a head-mounted-display system. The special-purpose hardware device may also include an input-formatting component that receives the computer-generated imagery. The special-purpose hardware device may further include a blending component that generates at least one mixed-reality frame by overlaying the computer-generated imagery onto the image frame. The special-purpose hardware device may additionally include a frame-output interface that feeds the mixed-reality frame generated by the blending component to a display device of the head-mounted-display system to facilitate displaying the mixed-reality frame for presentation to a user wearing the head-mounted-display system. Various other apparatuses, systems, and methods are also disclosed.

Abstract: The disclosed special-purpose hardware device may include an image signal processor that receives, from a camera device of a head-mounted-display system, image frames of a physical environment. The special-purpose hardware device may also include a positional tracking component that (1) stores at least a portion of the image frames in a cache of the special-purpose hardware device that has a faster access speed than a main memory of the special-purpose hardware device, (2) tracks, based on the portion of the image frames stored in the cache, a change in the position of the head-mounted display system within the physical environment, and (3) stores the change in the position of the head-mounted-display system in the main memory for use in generating one or more augmented-reality frames. The special-purpose hardware device may further include a frame-output interface that feeds the augmented-reality frames to a display device of the head-mounted-display system.

Abstract: A special-purpose hardware device for mitigating motion-to-photon latency in head-mounted displays may include an image signal processor that receives at least one image frame captured by a camera device of a head-mounted-display system. The special-purpose hardware device may also include an input-formatting component that receives the computer-generated imagery. The special-purpose hardware device may further include a blending component that generates at least one mixed-reality frame by overlaying the computer-generated imagery onto the image frame. The special-purpose hardware device may additionally include a frame-output interface that feeds the mixed-reality frame generated by the blending component to a display device of the head-mounted-display system to facilitate displaying the mixed-reality frame for presentation to a user wearing the head-mounted-display system. Various other apparatuses, systems, and methods are also disclosed.

Abstract: A system, method, and computer program product for improving digital images and video by reducing or eliminating artifacts created by sensor binning, i.e. channel displacement of downscaled image pixels. Red and blue pixels are shifted to corresponding ideal pixel locations using for example bi-linear interpolation for each color. Green pixels are shifted to corresponding ideal pixel locations using one-dimensional cubic interpolation along a diagonal direction in which the green pixels are aligned. Pixel values are replaced by weighted pixel value averages of groups of pixels, preferably four, and the weights used vary inversely with shifted distances. Cubic interpolation results may be separately weighted pairwise among pixels with a weighting parameter favoring pixels near the ideal pixel location. The embodiments noticeably improve image and video quality, particularly by treatment of jagged edges of diagonal image features without compromising image sharpness or creating false colors along edges.

Abstract: Techniques for implementing a Black Level Correction (BLC) processing operation on image data signal pixel values that results in little to no nonlinearity in the dark areas of the image due to black noise clipping, and avoids reducing image quality or adding cost, are provided. Image data signal pixel values are caused to maintain black level while being operated on by image data signal processor circuits that precede a Noise Reduction (NR) processing operation, thus allowing the BLC processing operation to be executed after the NR processing operation. With black noise mostly removed, little to no nonlinearity in the dark areas of the image due to black noise clipping results from the BLC processing operation.

Abstract: Techniques for implementing a Black Level Correction (BLC) processing operation on image data signal pixel values that results in little to no nonlinearity in the dark areas of the image due to black noise clipping, and avoids reducing image quality or adding cost, are provided. Image data signal pixel values are caused to maintain black level while being operated on by image data signal processor circuits that precede a Noise Reduction (NR) processing operation, thus allowing the BLC processing operation to be executed after the NR processing operation. With black noise mostly removed, little to no nonlinearity in the dark areas of the image due to black noise clipping results from the BLC processing operation.