Nvidia Corporation

Sean Midthun Pieper

Robin Brian Jenkin

Animesh Khemka

Ricardo Jansson Motta

Eric Dujardin

Brian Keith Cabral

Minwoo Park

Yining Deng

Ross Cunniff

Seungseok Oh

Kurt Roland Wall

Amit Purwar

Michael Stevens

A method of compression includes compressing a captured image from a first bit-depth at which it was captured to a second bit-depth that is less than the first bit-depth. The compression comprises applying, according to a power curve, a first compression amount to a first region of the captured image having a first pixel value; and applying, according to the power curve, a second compression amount to a second part of the captured image having a higher second pixel value, wherein the first compression amount is lower than the second compression amount.

Image signal processing pipelines for high dynamic range sensors

In one embodiment, a system receives pixel data from a pair of regions of an image generated by an imaging device, the pair of regions includes a first region and a second region, where the first region includes a first plurality of pixels and the second region includes a second plurality of pixels. The system determines a plurality of pixel pairs, where a pixel pair includes a first pixel from the first plurality of pixels and a second pixel from the second plurality of pixels. The system calculates a plurality of contrasts based on the plurality of pixel pairs. The system determines a contrast distribution based on the plurality of contrasts. The system calculates a value representative of a capability of the imaging device to detect contrast based on the contrast distribution. The system determines a reduction in contrast detectability of the imaging device based on the value.

Applications for detection capabilities of cameras

In various examples, apparatuses, systems, and techniques to perform offline image signal processing of source image data to generate target image data. In at least one embodiment, data collection using exposure and calibration setting of an image sensor is performed to generate source image data, which is then processed by using offline image signal processing to generate target data.

High dynamic range image processing with fixed calibration settings

Apparatuses, systems, and techniques to receive, at one or more processors associated with an image signal processing (ISP) pipeline for a camera, an image generated using an image sensor of the camera, wherein the image comprises a plurality of channels associated with color information of the image; process, by the one or more processors, the plurality of channels of the image to generate a plurality of luminance and/or radiance values; generate, by the one or more processors, an updated version of the image using the plurality of luminance and/or radiance values; and output the updated version of the image.

Determination of luminance values using image signal processing pipeline

Apparatuses, systems, and techniques to process luminance and/or radiance values of one or more images from one or more cameras using one or more neural networks to perform a machine vision task. In at least one embodiment, one or more neural networks determine detection difficulty levels of objects within the one or more images and performs a machine vision task based on the determined detection difficulty levels of objects within images associated with that ask.

Training a neural network using luminance

A method of decompression includes decompressing a compressed image according to a power curve to generate a partially decompressed image, wherein the compressed image is decompressed from a second bit depth that is lower than a first bit-depth at which the image was generated.

Apparatuses, systems, and techniques to optimize memory usage when performing matrix operations. In at least one embodiment, a matrix is optimized to limit memory and storage requirements while minimizing loss of precision for a sum of the members of the matrix.

Kernel coefficient quantization

In one embodiment, a system determines pixel data from a pair of regions of an image generated by an imaging device, the pair of regions includes a first region and a second region, where the first region includes a first plurality of pixels and the second region includes a second plurality of pixels. The system determines a plurality of pixel pairs of the image, where a pixel pair includes a first pixel from the first plurality of pixels and a second pixel from the second plurality of pixels. The system calculates a plurality of contrasts based on the plurality of pixel pairs, where a contrast is calculated between the first pixel and the second pixel. The system determines a contrast distribution based on the plurality of contrasts. The system calculates a value representative of a capability of the imaging device to detect contrast based on the contrast distribution.

Evaluating detection capabilities of cameras

Apparatuses, systems, and techniques to receive, at one or more processor associated with an image signal processing (ISP) pipeline, a compressed image generated by an image sensor, wherein the compressed image is captured at a first bit-depth associated with the image sensor and is compressed to a second bit-depth that is lower than the first bit-depth, and wherein the ISP is associated with a third bit-depth that is lower than the first bit-depth and higher than the second bit-depth; and decompress the compressed image according to a power curve to generate a partially decompressed image having the third bit-depth, wherein a plurality of regions of the partially decompressed image are decompressed at separate decompression amounts based on a corresponding pixel value of each region of the plurality of regions.

Systems and techniques for noise reduction in video are described. Example implementations provide improved motion-adaptive temporal or spatio-temporal noise reduction that use an improved blending of the current frame with previous frames. The improved blending may be particularly effective for processing video captured in noisy environments such as low-light and/or mobile environments. In some example implementations, the improved blending is based on more accurately distinguishing between pixel difference in adjacent images that are caused by motion rather than noise.

Techniques for reducing noise in video

Systems and methods for generating high dynamic images from interleaved Bayer array data with high spatial resolution and reduced sampling artifacts. Bayer array data are demosaiced into components of the YUV color space before deinterleaving. The Y component and the UV components can be derived from the Bayer array data through demosiac convolution processes. A respective convolution is performed between a convolution kernel and a set of adjacent pixels of the Bayer array that are in the same color channel. A convolution kernel is selected based the mosaic pattern of the Bayer array and the color channels of the set of adjacent pixels. The Y data and UV data are deinterleaved and interpolated into frames of short exposure and long exposures in the second color space. The short exposure and long exposure frames are then blended and converted back to a RGB frame representing a high dynamic range image.

Deinterleaving interleaved high dynamic range image by using YUV interpolation

Dynamic white point management techniques include determining a white point of ambient light proximate to a display. A color profile adjustment is determined based upon the determined white point and intensity of the ambient light. The image color space is transformed to a display color space for rendering on the display based on the determined adjusted to the color profile.

Dynamic color profile management for electronic devices

Embodiments of the present invention are directed to methods and systems for robust weighting of gray patches in automatic white balancing in an image-capture device by utilizing kernel density estimation techniques with dynamically variable bandwidth to determine the probability density of samples to create an initial estimate, then verifying the initial gray point estimate to account for outliers. In one embodiment, given a set of image data, an initial gray point estimate in a color space is determined for the set of image data. The initial estimate is then refined by weighting the sub-population with the greatest probability of being gray. A final evaluation that includes a further comparison to pre-programmed constraints determines a final estimate, which can still be further tuned according to user preferences by adjusting color biases. The resulting final gray point estimate provides greater stability, and greatly improved accuracy over traditional techniques and solutions.

Growing community of inventors

San Jose, CA, United States of America

Sean Midthun Pieper