Abstract: The present disclosure provides techniques relates to the implementation of a raw pixel processing unit using a set of line buffers. In one embodiment, the set of line buffers may include a first subset and second subset. Various logical units of the raw pixel processing unit may be implemented using the first and second subsets of line buffers in a shared manner. For instance, in one embodiment, defective pixel correction and detection logic may be implemented using the first subset of line buffers. The second subset of line buffers may be used to implement lens shading correction logic, gain, offset, and clamping logic, and demosaicing logic. Further, noise reduction may also be implemented using at least a portion of each of the first and second subsets of line buffers.

Abstract: Various techniques are provided for processing image data acquired using a digital image sensor. In accordance with aspects of the present disclosure, one such technique may relate to the processing of image data in a system that supports multiple image sensors. In one embodiment, the image processing system may include control circuitry configured to determine whether a device is operating in a single sensor mode (one active sensor) or a dual sensor mode (two active sensors). When operating in the single sensor mode, data may be provided directly to a front-end pixel processing unit from the sensor interface of the active sensor. When operating in a dual sensor mode, the image frames from the first and second sensors are provided to the front-end pixel processing unit in an interleaved manner. For instance, in one embodiment, the image frames from the first and second sensors are written to a memory, and then read out to the front-end pixel processing unit in an interleaved manner.

Abstract: Certain aspects of this disclosure relate to an image signal processing system that includes a flash controller that is configured to activate a flash device prior to the start of a target image frame by using a sensor timing signal. In one embodiment, the flash controller receives a delayed sensor timing signal and determines a flash activation start time by using the delayed sensor timing signal to identify a time corresponding to the end of the previous frame, increasing that time by a vertical blanking time, and then subtracting a first offset to compensate for delay between the sensor timing signal and the delayed sensor timing signal. Then, the flash controller subtracts a second offset to determine the flash activation time, thus ensuring that the flash is activated prior to receiving the first pixel of the target frame.

Abstract: Various techniques for lens shading correction are provided. In one embodiment, the location of a current pixel is determined relative to a gain grid having a plurality of grid points distributed in horizontal and vertical directions. If the location of the current pixel corresponds to a grid point, a lens shading gain associated with that grid point is applied to the current pixel. If the location of the current pixel is between four grid points, bi-linear interpolation is applied to the four grid points to determine an interpolated lens shading gain. In another embodiment, a radial gain grid may be provided, and lens shading gains may be interpolated based upon grid points neighboring a current pixel in the radial and angular directions. In a further embodiment, a radial lens shading gain is determined by determining a radial distance from the center of the image to the current pixel and multiplying the radial distance by a global gain parameter based upon the color of the current pixel.

Abstract: Systems and methods are disclosed for applying spatial filtering to raw image data. For example, a spatial filter may identify an n x n block of pixels from an image frame. The spatial filter has filter taps, each corresponding to one of the pixels of the n x n block. Pixel difference values between the input pixel and each of the neighboring pixels in the n x n block may be determined. Attenuation factors may be determined based on the pixel differences and brightness of the input pixel. Applying the attenuation factors to their respective filtering taps may produce attenuated filtering coefficients that are used to obtain filtered pixel values. By normalizing the sum of the filtered pixel values using the sum of the attenuated filtering coefficients, a spatially filtered output value corresponding to the current input pixel (e.g., located at the center of the n x n block) may be determined.

Abstract: A method for encoding a picture is disclosed. The method generally includes the steps of (A) generating at least one respective macroblock statistic from each of a plurality of macroblocks in the picture, (B) generating at least one global statistic from the picture and (C) generating a respective macroblock quantization parameter for each of the macroblocks based on both (i) the at least one respective macroblock statistic and (ii) said at least one global statistic.

Abstract: A device, method, computer useable medium, and processor programmed to automatically generate tone mapping curves in a digital camera based on image metadata are described. By examining image metadata from a digital camera's sensor, such as the light-product, one can detect sun-lit, high-light, and low-light scenes. Once the light-product value has been calculated for a given image, a tone mapping curve can automatically be generated within the sensor and adjusted appropriately for the scene based on predetermined parameters. Further, it has been determined that independently varying the slopes of the tone mapping curve at the low end (S0) and high end (S1) of the curve results in more visually appealing images. By dynamically and independently selecting S0 and S1 values based on image metadata, more visually pleasing images can be generated.

Abstract: Various techniques are provided herein for processing raw image data acquired using a digital image sensor in an image processing pipeline of an image signal processing system. In one embodiment, the image processing pipeline may first process the raw image data (e.g., Bayer image data) for the detection and correction of defective pixels. Next, the image processing pipeline may process the raw image data to reduce noise. Thereafter, the image processing pipeline may correct lens shading distortion in the raw image data and, subsequently, apply a demosaicing algorithm to convert the raw image data into full color image data (e.g., RGB image data). The color image data may be further processed by the image processing pipeline to correct color and gamma properties prior to being converted into a luma and chroma color space (e.g., YCbCr color space).

Abstract: A camera comprising a first circuit and a second circuit. The first circuit may be configured to perform image signal processing using encoding related information. The second circuit may be configured to encode image data using image signal processing related information. The first circuit may be further configured to pass the image signal processing related information to the second circuit. The second circuit may be further configured to pass the encoding related information to the first circuit. The second circuit may be further configured to modify one or more motion estimation processes based upon the information from the first circuit.

Abstract: Various techniques are provided herein for processing raw image data in front-end processing logic of an image signal processing system. In one embodiment, the front-end processing logic includes a statistics processing unit configured to process raw image data acquired by an image sensor to obtain one or more sets of statistics. The statistics processing unit may first correct defective pixels in the raw image data and then correct lens shading errors in the raw image data prior to extracting the statistics information. In certain embodiments, black level compensation may be applied between the defective pixel correction and lens shading correction steps, and inverse black level compensation may be applied between the lens shading correction step and the extraction of the statistics information. The acquired statistics information may be utilized by an image signal processing pipeline for converting the raw image data into a color (e.g., RGB) and/or luma (e.g., YCbCr) image.

Abstract: A method for encoding an image is disclosed. The method generally includes the steps, of (A) generating a quantization matrix as a function of at least four parameters, (B) optimizing the parameters to maximize a quality metric for encoding the image and (C) encoding the image with the quantization matrix as optimized.

Abstract: Various techniques relating to image sharpening are provided. In one embodiment, a luminance image is obtained based upon image data acquired by an image sensor. A multi-scale unsharp mask, which may include at least two Gaussian filters of difference radii, is applied to the luminance image to determine a plurality of unsharp values. Each of the unsharp values may be compared to a corresponding threshold and, for the unsharp values that exceed their respective thresholds, the unsharp value is multiplied by a corresponding gain and added to a base image, which may be selected as the luminance image or the output of one of the Gaussian filters. Each gained unsharp value may be summed with the base image to produce a final sharpened output. In some embodiments, an attenuated gain may be applied to unsharp values that do not exceed their respective thresholds.

Abstract: A method to control weighted prediction for video compression. The method comprises (A) generating statistics based upon analysis of a plurality of video frames, (B) detecting a fade condition based upon the statistics, (C) generating one or more estimated weighted prediction parameters based upon the statistics when the fade condition is detected and (D) encoding the plurality of video frames. The encoding takes into account the estimated weighted prediction parameters when generated.

Abstract: A video transcoder is disclosed. The video transcoder generally comprises a processor and a video digital signal processor. The processor may be formed on a first die. The video digital signal processor may be formed on a second die and coupled to the processor. The video digital signal processor may have (i) a first module configured to perform a first operation in decoding an input video stream in a first format and (ii) a second module configured to perform a second operation in coding an output video stream in a second format, wherein the first operation and the second operation are performed in parallel.

Abstract: A method for transcoding from an MPEG-2 format to an H.264 format is disclosed. The method generally comprises the steps of (A) decoding an input video stream in the MPEG-2 format to generate a plurality of macroblocks; (B) determining a plurality of indicators from a pair of the macroblocks, the pair of the macroblocks being vertically adjoining; and (C) coding the pair of the macroblocks into an output video stream in the H.264 format using one of (i) a field mode coding and (ii) a frame mode coding as determined from the indicators.

Abstract: Various techniques are provided for the detection and correction of defective pixels in an image sensor. In accordance with one embodiment, a static defect table storing the locations of known static defects is provided, and the location of a current pixel is compared to the static defect table. If the location of the current pixel is found in the static defect table, the current pixel is identified as a static defect and is corrected using the value of the previous pixel of the same color. If the current pixel is not identified as a static defect, a dynamic defect detection process includes comparing pixel-to-pixel gradients between the current pixel a set of neighboring pixels against a dynamic defect threshold. If a dynamic defect is detected, a replacement value for correcting the dynamic defect may be determined by interpolating the value of two neighboring pixels on opposite sides of the current pixel in a direction exhibiting the smallest gradient.

Abstract: A method for deinterlacing a picture is disclosed. The method generally includes the steps of (A) generating a plurality of primary scores by searching along a plurality of primary angles for an edge in the picture proximate a location interlaced with a field of the picture, (B) generating a plurality of neighbor scores by searching for the edge along a plurality of neighbor angles proximate a particular angle of the primary angles corresponding to a particular score of the primary scores having a best value and (C) identifying a best score from a group of scores consisting of the particular score and the neighbor scores to generate an interpolated sample at the location.

Abstract: A video transcoder is disclosed. The video transcoder generally comprises a processor and a video digital signal processor. The processor may be formed on a first die. The video digital signal processor may be formed on a second die and coupled to the processor. The video digital signal processor may have (i) a first module configured to perform a first operation in decoding an input video stream in a first format and (ii) a second module configured to perform a second operation in coding an output video stream in a second format, wherein the first operation and the second operation are performed in parallel.

Abstract: A method for transcoding from an H.264 format to an MPEG-2 format is disclosed. The method generally comprises the steps of (A) decoding an input video stream in the H.264 format to generate a picture having a plurality of macroblock pairs that used an H.264 macroblock adaptive field/frame coding; (B) determining a mode indicator for each of the macroblock pairs; and (C) coding the macroblock pairs into an output video stream in the MPEG-2 format using one of (i) an MPEG-2 field mode coding and (ii) an MPEG-2 frame mode coding as determined from the mode indicators.

Abstract: An apparatus comprising a first circuit, a second circuit and a third circuit. The first circuit may be configured to determine one or more signal characteristics in a portion of an input video signal. The second circuit may be configured to select a multiplier value from a plurality of multiplier values in response to the signal characteristics. The third circuit may be configured to generate an encoded bitstream in response to (i) the input video signal and (ii) the selected multiplier value.