Abstract: A computer implemented method and apparatus for normalizing and predicting time series data. The method comprises accessing collected data comprising a plurality of intervals; defining a variation for each interval in the plurality of intervals, wherein each variation is a cost value; clustering the cost values, wherein clustering identifies anomalies in the collected data; correcting the anomalies in the collected data; and creating a set of normalized data from the corrected data.

Abstract: A method, system, and computer-readable storage medium are disclosed for generating fused depth maps. A plurality of initial depth maps are generated for a first image of a stereo pair, using at least two separate and distinct techniques for depth estimation. A plurality of initial depth maps are generated for a second image of the stereo pair, using at least two separate and distinct techniques for depth estimation. A fused depth map is generated for the first image based on the plurality of initial depth maps for the first image and the plurality of initial depth maps for the second image. A fused depth map is generated for the second image based on the plurality of initial depth maps for the second image and the plurality of initial depth maps for the first image.

Abstract: A computer implemented method and apparatus for normalizing and predicting time series data. The method comprises accessing collected data comprising a plurality of intervals; defining a variation for each interval in the plurality of intervals, wherein each variation is a cost value; clustering the cost values, wherein clustering identifies anomalies in the collected data; correcting the anomalies in the collected data; and creating a set of normalized data from the corrected data.

Abstract: Methods and apparatus for deinterlacing an interlaced image or for upsampling an image. In one embodiment, an Edge Aware Deinterlacing system may identify edge areas and non-edge areas of an image in order to apply one interpolation method to pixels within the edge areas and a different interpolation method to pixels within the non-edge areas.

Abstract: Various embodiments of methods and apparatus for fitting a surface to a data set are disclosed. A frequency distribution of an input data set is determined. Determining the frequency distribution includes assigning each data point of the input data set to a category representing a value of a variable for the respective data point. Responsive to identifying one or more discontinuities of the frequency distribution, a continuous section of the frequency distribution is identified as a first data set. A first equation is fit to the first data set.

Abstract: An optimized method and system for entropy coding is described. The system may include an input module to receive a digital image, a Discrete Cosine Transform (DCT) module to obtain a plurality of DCT coefficients associated with the digital image, a statistics generator to generate statistics associated with the plurality of DCT coefficients, a placement module to store the generated statistics in the plurality of DCT coefficients and entropy coder to generate uniquely decodable (UD) codes for the plurality of DCT coefficients. The entropy coder may utilize the statistics stored in the plurality of DCT coefficients.

Abstract: A system that compresses a digital image. During operation, the system obtains a digital image in which the image components are de-correlated into separate channels containing luminance information and chrominance information. Next, the system receives a quality number for the luminance channel and a quality number for the chrominance channels, and uses the quality numbers to determine a target bitrate for the luminance channel and for the chrominance channels, respectively. Given the target bitrate for the luminance channel and the target bit rate for the chrominance channels, the system performs rate distortion optimizations separately on the luminance and chrominance channels to determine which portions of the luminance bitstream and which portions of the chrominance bitstream are to be included in the final bitstream. The system then uses the stored rate-distortion information to generate a preview of the digital image without performing entropy coding each time the quality settings are changed.

Abstract: A method, system, and computer-readable storage medium are disclosed for generating fused depth maps. A plurality of initial depth maps are generated for a first image of a stereo pair, using at least two separate and distinct techniques for depth estimation. A plurality of initial depth maps are generated for a second image of the stereo pair, using at least two separate and distinct techniques for depth estimation. A fused depth map is generated for the first image based on the plurality of initial depth maps for the first image and the plurality of initial depth maps for the second image. A fused depth map is generated for the second image based on the plurality of initial depth maps for the second image and the plurality of initial depth maps for the first image.

Abstract: Various embodiments of methods and apparatus for fitting a surface to a data set are disclosed. A frequency distribution of an input data set is determined. Determining the frequency distribution includes assigning each data point of the input data set to a category representing a value of a variable for the respective data point. Responsive to identifying one or more discontinuities of the frequency distribution, a continuous section of the frequency distribution is identified as a first data set. A first equation is fit to the first data set.

Abstract: Methods and apparatus for deinterlacing an interlaced image or for upsampling an image. In one embodiment, an Edge Aware Deinterlacing system may identify edge areas and non-edge areas of an image in order to apply one interpolation method to pixels within the edge areas and a different interpolation method to pixels within the non-edge areas.

Abstract: Embodiments include a computer implemented method and apparatus for normalizing perceptual quality in media compression. In one embodiment, a computer implemented method for normalizing perceptual quality in media compression includes determining a maximum bit rate for a target media file, determining a quality setting for the target media file, and compressing the media file to a bit rate equal to the specific percentage of the maximum bit rate multiplied by the maximum bit rate using the computer. The quality setting is a specific percentage of the maximum bit rate.

Abstract: An optimized method and system for entropy coding is described. The system may include an input module to receive a digital image, a Discrete Cosine Transform (DCT) module to obtain a plurality of DCT coefficients associated with the digital image, a statistics generator to generate statistics associated with the plurality of DCT coefficients, a placement module to store the generated statistics in the plurality of DCT coefficients and entropy coder to generate uniquely decodable (UD) codes for the plurality of DCT coefficients. The entropy coder may utilize the statistics stored in the plurality of DCT coefficients.

Abstract: An optimized method and system to generate a compressed image utilizing custom probability tables are described. The system may include a input module to receive a digital image, a sample generator to generate a representative sample image from the input image, a custom probability tables generator to generate custom probability tables and an entropy coder to encode the input image utilizing the custom probability tables. The custom probability tables generator may be configured to generate custom probability tables utilizing the representative sample image.

Abstract: An optimized method and system for entropy coding is described. The system may include an input module to receive a digital image, a Discrete Cosine Transform (DCT) module to obtain a plurality of DCT coefficients associated with the digital image, a statistics generator to generate statistics associated with the plurality of DCT coefficients, a placement module to store the generated statistics in the plurality of DCT coefficients and entropy coder to generate uniquely decodable (UD) codes for the plurality of DCT coefficients. The entropy coder may utilize the statistics stored in the plurality of DCT coefficients.

Abstract: Systems, methods, and apparatus, including computer program products, are provided for changing the look of an image. In some implementations a computer-implemented method is provided. The method includes decomposing a first image using a first plurality of wavelet transforms, each first wavelet transform creating a first transformation resolution, each first transformation resolution having respective first data, and decomposing a second image using a second plurality of wavelet transforms, each second wavelet transform creating a second transformation resolution, each second transformation resolution having respective second data. The method also includes determining a first measure of texture for each first transformation resolution's respective data, applying each first transformation resolution's first measure of texture to each corresponding second transformation resolution's data, and recomposing the second image based on the modified second transformation resolutions.

Abstract: Methods and apparatus for localized labeling in digital images. A region is obtained within which a global labeling solution for an image lies. The region is covered with a set of multiple overlapping tiles. A labeling function is applied to each tile in two or more subsets of the tiles to generate a local labeling for each of the tiles in the subsets. The local labeling for tiles in a first subset are input as a boundary condition to the labeling function when applied to overlapping tiles in a second subset. The local labelings for all of the tiles in all of the subsets are merged to form a global labeling for the image. The labeling function may be executed in parallel on two or more of the tiles.

Abstract: A system that compresses a digital image. During operation, the system obtains a digital image in which the image components are de-correlated into separate channels containing luminance information and chrominance information. Next, the system receives a quality number for the luminance channel and a quality number for the chrominance channels, and uses the quality numbers to determine a target bitrate for the luminance channel and for the chrominance channels, respectively. Given the target bitrate for the luminance channel and the target bit rate for the chrominance channels, the system performs rate distortion optimizations separately on the luminance and chrominance channels to determine which portions of the luminance bitstream and which portions of the chrominance bitstream are to be included in the final bitstream. The system then uses the stored rate-distortion information to generate a preview of the digital image without performing entropy coding each time the quality settings are changed.

Abstract: A system that compresses a digital image. During operation, the system obtains a digital image in which the image components are de-correlated into separate channels containing luminance information and chrominance information. Next, the system receives a quality number for the luminance channel and a quality number for the chrominance channels, and uses the quality numbers to determine a target bitrate for the luminance channel and for the chrominance channels, respectively. Given the target bitrate for the luminance channel and the target bit rate for the chrominance channels, the system performs rate distortion optimizations separately on the luminance and chrominance channels to determine which portions of the luminance bitstream and which portions of the chrominance bitstream are to be included in the final bitstream. The system then uses the stored rate-distortion information to generate a preview of the digital image without performing entropy coding each time the quality settings are changed.

Abstract: Methods and apparatus, including computer program products, for compression include decomposing an image into wavelet transform coefficients, applying a first quantizer to the wavelet transform coefficients, and applying a second quantizer to a set of wavelet transform coefficients representing a background of the image. The method can include rounding the wavelet transform coefficients and rounding the wavelet transform coefficients can include rounding to a nearest integer value. The method can include multiplying the set rounded quantized background wavelet transform coefficients by the second quantizer and rounding all wavelet transform coefficients. Rounding all wavelet transform coefficients can include rounding to a nearest integer value.

Abstract: Methods and apparatus implementing systems and techniques for determining quantizer values for a signal. In general, in one implementation, a technique includes: obtaining a target distortion and a transformed signal having sub-bands resulting from a source signal being transformed by a wavelet-based decomposition; distributing the target distortion among the sub-bands based on energy weights to produce target sub-band distortions; ascertaining actual sub-band distortion values that result at different grades of quantization; determining quantizer values for the sub-bands, including interpolating between at least two of the quantization grades; and quantizing the sub-bands with the determined quantizer values. The signal may be a digital image. Additionally, in another implementation, determining the quantizer values can involve determining approximations of sub-band coefficient quantization error rather than performing the interpolation, in which case, the ascertaining operation need not be performed.