公开(公告)号：US20190356916A1

公开(公告)日：2019-11-21

申请号：US15985317

申请日：2018-05-21

Applicant: Google LLC

Inventor： Krzysztof Potempa ,  Jyrki Alakuijala ,  Robert Obryk

IPC: H04N19/122 ,  H04N19/176 ,  H04N19/625

Abstract: An image encoder includes a processor and a memory. The memory includes instructions configured to cause the processor to perform operations. In one example implementation, the operations may include determining whether a dictionary item is available for replacing a block of an image being encoded, the determining based on a hierarchical lookup mechanism, and encoding the image along with reference information of the dictionary item in response to determining that the dictionary item is available. In one more example implementation, the operations may include performing principal component analysis (PCA) on a block to generate a corresponding projected block, the block being associated with a group of images, comparing the projected block with a corresponding threshold, descending the block recursively based on the threshold until a condition is satisfied, and identifying a left over block as a cluster upon satisfying of the condition.

公开(公告)号：US11968406B2

公开(公告)日：2024-04-23

申请号：US17248795

申请日：2021-02-08

Applicant: Google LLC

Inventor： Krzysztof Potempa ,  Jyrki Alakuijala ,  Robert Obryk

IPC: H04N19/85 ,  G06F18/231 ,  G06V10/762 ,  G06V10/77 ,  G06V10/772 ,  H03M7/30 ,  H04N19/122 ,  H04N19/176 ,  H04N19/625

CPC classification number: H04N19/85 ,  G06F18/231 ,  G06V10/7625 ,  G06V10/7715 ,  G06V10/772 ,  H03M7/3088 ,  H04N19/122 ,  H04N19/176 ,  H04N19/625 ,  H03M7/3077

Abstract: An image encoder includes a processor and a memory. The memory includes instructions configured to cause the processor to perform operations. In one example implementation, the operations may include determining whether a dictionary item is available for replacing a block of an image being encoded, the determining based on a hierarchical lookup mechanism, and encoding the image along with reference information of the dictionary item in response to determining that the dictionary item is available. In one more example implementation, the operations may include performing principal component analysis (PCA) on a block to generate a corresponding projected block, the block being associated with a group of images, comparing the projected block with a corresponding threshold, descending the block recursively based on the threshold until a condition is satisfied, and identifying a left over block as a cluster upon satisfying of the condition.

公开(公告)号：US20220405058A1

公开(公告)日：2022-12-22

申请号：US17821979

申请日：2022-08-24

Applicant: Google LLC

Inventor： Jan Wassenberg ,  Robert Obryk ,  Jyrki Alakuijala ,  Emmanuel Mogenet

IPC: G06F7/58 ,  H04L9/06 ,  H04L9/08

Abstract: A method for generating random numbers includes initializing a pseudo-random number generator (PRNG) having a state of 2048 bits comprising inner bits and outer bits, the inner bits comprising the first 128 bits of the 2048 bits and the outer bits comprising the remaining bits of the 2048 bits. The method also includes retrieving AES round keys from a key source, and for a threshold number of times, executing a round function using the AES round keys by XOR'ing odd-numbered branches of a Feistel network having 16 branches of 128 bits with a function of corresponding even-numbered neighbor branches of the Feistel network, and shuffling each branch of 128 bits into a prescribed order. The method also includes executing an XOR of the inner bits of the permuted state with the inner bits of a previous state.

公开(公告)号：US11425281B1

公开(公告)日：2022-08-23

申请号：US15258368

申请日：2016-09-07

Applicant: GOOGLE LLC

Inventor： Robert Obryk ,  Jyrki Antero Alakuijala

IPC: G06K9/00 ,  H04N1/60 ,  H04N1/64 ,  G06V10/28

Abstract: Techniques of color image processing involve performing a transformation for each color channel that mixes intensity values from other channels to produce a new intensity value for that channel. The new intensity values, representing the effect of overlapped response spectra of the S, M, and L cones, then provide values of the sensitivities of the photoreceptors of each of the cones. These values of the sensitivities form the basis of more accurate color image processing. For example, compression ratios of gamma-compressed color images may be increased when more the sensitivities are more accurate.

公开(公告)号：US20240276018A1

公开(公告)日：2024-08-15

申请号：US18643085

申请日：2024-04-23

Applicant: GOOGLE LLC

Inventor： Jyrki Alakuijala ,  Robert Obryk ,  Evgenii Kliuchnikov ,  Zoltan Szabadka ,  Jan Wassenberg ,  Minttu Alakuijala ,  Lode Vandevenne

IPC: H04N19/65 ,  H04N19/124 ,  H04N19/18 ,  H04N19/184 ,  H04N19/85

CPC classification number: H04N19/65 ,  H04N19/124 ,  H04N19/18 ,  H04N19/184 ,  H04N19/85

Abstract: The loss of image quality during compression is controlled using a sequence of quality control metrics. The sequence of quality control metrics is selected for quantizing transform coefficients within an area of the image based on an error level definition. Candidate bit costs are then determined by quantizing the transform coefficients according to the error level definition or a modified error level and the sequence of quality control metrics. Where the candidate bit cost resulting from using the modified error level is lower than the candidate bit cost resulting from using the error level definition, the transform coefficients are quantized according to the modified error level and the sequence of quality control metrics. Otherwise, the transform coefficients are quantized based on the error level definition and according to the sequence of quality control metrics.

公开(公告)号：US11734568B2

公开(公告)日：2023-08-22

申请号：US16274599

申请日：2019-02-13

Applicant: Google LLC

Inventor： Jyrki Alakuijala ,  Ruud van Asseldonk ,  Robert Obryk ,  Krzysztof Potempa

IPC: G06N3/082 ,  G06N3/084 ,  G06N20/20 ,  G06N3/045

CPC classification number: G06N3/082 ,  G06N3/045 ,  G06N3/084 ,  G06N20/20

Abstract: The present disclosure provides systems and methods for modification (e.g., pruning, compression, quantization, etc.) of artificial neural networks based on estimations of the utility of network connections (also known as “edges”). In particular, the present disclosure provides novel techniques for estimating the utility of one or more edges of a neural network in a fashion that requires far less expenditure of resources than calculation of the actual utility. Based on these estimated edge utilities, a computing system can make intelligent decisions regarding network pruning, network quantization, or other modifications to a neural network. In particular, these modifications can reduce resource requirements associated with the neural network. By making these decisions with knowledge of and based on the utility of various edges, this reduction in resource requirements can be achieved with only a minimal, if any, degradation of network performance (e.g., prediction accuracy).

公开(公告)号：US11755287B2

公开(公告)日：2023-09-12

申请号：US17821979

申请日：2022-08-24

Applicant: Google LLC

Inventor： Jan Wassenberg ,  Robert Obryk ,  Jyrki Alakuijala ,  Emmanuel Mogenet

IPC: H04L29/06 ,  G06F21/00 ,  G06F7/58 ,  H04L9/06 ,  H04L9/08

CPC classification number: G06F7/582 ,  H04L9/0625 ,  H04L9/0631 ,  H04L9/0869

Abstract: A method for generating random numbers includes initializing a pseudo-random number generator (PRNG) having a state of 2048 bits comprising inner bits and outer bits, the inner bits comprising the first 128 bits of the 2048 bits and the outer bits comprising the remaining bits of the 2048 bits. The method also includes retrieving AES round keys from a key source, and for a threshold number of times, executing a round function using the AES round keys by XOR'ing odd-numbered branches of a Feistel network having 16 branches of 128 bits with a function of corresponding even-numbered neighbor branches of the Feistel network, and shuffling each branch of 128 bits into a prescribed order. The method also includes executing an XOR of the inner bits of the permuted state with the inner bits of a previous state.

公开(公告)号：US11531932B2

公开(公告)日：2022-12-20

申请号：US16624497

申请日：2017-07-06

Applicant: Google LLC

Inventor： Jyrki Alakuijala ,  Robert Obryk

IPC: G06N20/00 ,  G06N3/04 ,  G06N3/08

Abstract: The present disclosure provides systems and methods for compressing and/or distributing machine learning models. In one example, a computer-implemented method is provided to compress machine-learned models, which includes obtaining, by one or more computing devices, a machine-learned model. The method includes selecting, by the one or more computing devices, a weight to be quantized and quantizing, by the one or more computing devices, the weight. The method includes propagating, by the one or more computing devices, at least a part of a quantization error to one or more non-quantized weights and quantizing, by the one or more computing devices, one or more of the non-quantized weights. The method includes providing, by the one or more computing devices, a quantized machine-learned model.

公开(公告)号：US11463733B2

公开(公告)日：2022-10-04

申请号：US16970499

申请日：2019-02-15

Applicant: Google LLC

Inventor： Jyrki Alakuijala ,  Robert Obryk ,  Evgenii Kliuchnikov ,  Zoltan Szabadka ,  Jan Wassenberg ,  Minttu Alakuijala ,  Lode Vandevenne

IPC: H04N19/65 ,  H04N19/124 ,  H04N19/18 ,  H04N19/184 ,  H04N19/85

Abstract: The loss of image quality during compression is controlled using a sequence of quality control metrics. The sequence of quality control metrics is selected for quantizing transform coefficients within an area of the image based on an error level definition. Candidate bit costs are then determined by quantizing the transform coefficients according to the error level definition or a modified error level and the sequence of quality control metrics. Where the candidate bit cost resulting from using the modified error level is lower than the candidate bit cost resulting from using the error level definition, the transform coefficients are quantized according to the modified error level and the sequence of quality control metrics. Otherwise, the transform coefficients are quantized based on the error level definition and according to the sequence of quality control metrics.

公开(公告)号：US11449311B2

公开(公告)日：2022-09-20

申请号：US16613516

申请日：2017-11-07

Applicant: Google LLC

Inventor： Jan Wassenberg ,  Robert Obryk ,  Jyrki Alakuijala ,  Emmanuel Mogenet

IPC: G06F21/00 ,  H04L29/06 ,  G06F7/58 ,  H04L9/06 ,  H04L9/08

Abstract: A method for generating random numbers includes initializing a pseudo-random number generator (PRNG) having a state of 2048 bits comprising inner bits and outer bits, the inner bits comprising the first 128 bits of the 2048 bits and the outer bits comprising the remaining bits of the 2048 bits. The method also includes retrieving AES round keys from a key source, and for a threshold number of times, executing a round function using the AES round keys by XOR'ing odd-numbered branches of a Feistel network having 16 branches of 128 bits with a function of corresponding even-numbered neighbor branches of the Feistel network, and shuffling each branch of 128 bits into a prescribed order. The method also includes executing an XOR of the inner bits of the permuted state with the inner bits of a previous state.