Patents by Inventor Akshay Gadde
Akshay Gadde has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Publication number: 20240137400Abstract: A media item to be provided to users of a platform is identified. The media item is associated with a media class of one or more media classes. An indication of the media item is provided as input to a machine learning model trained based on historical encoding data to predict, for a given media item, a set of encoder parameter settings that satisfy a performance criterion in view of a respective media class of the given media item. The historical encoding data includes a prior set of encoder parameter settings that satisfied the performance criterion with respect to a prior media item associated with the respective class. Encoder parameter settings that satisfy the performance criterion in view of the media class is determined based on an output of the model. The media item is caused to be encoded using the determined encoder parameter settings.Type: ApplicationFiled: January 3, 2024Publication date: April 25, 2024Inventors: Ching Yin Derek Pang, Kyrah Felder, Akshay Gadde, Paul Wilkins, Cheng Chen, Yao-Chung Lin
-
Patent number: 11924449Abstract: A learning model is trained for rate-distortion behavior prediction against a corpus of a video hosting platform and used to determine optimal bitrate allocations for video data given video content complexity across the corpus of the video hosting platform. Complexity features of the video data are processed using the learning model to determine a rate-distortion cluster prediction for the video data, and transcoding parameters for transcoding the video data are selected based on that prediction. The rate-distortion clusters are modeled during the training of the learning model, such as based on rate-distortion curves of video data of the corpus of the video hosting platform and based on classifications of such video data. This approach minimizes total corpus egress and/or storage while further maintaining uniformity in the delivered quality of videos by the video hosting platform.Type: GrantFiled: May 19, 2020Date of Patent: March 5, 2024Assignee: GOOGLE LLCInventors: Sam John, Balineedu Adsumilli, Akshay Gadde
-
Patent number: 11870833Abstract: Methods and systems for encoder parameter setting optimization. A media item to be provided to one or more users of a platform is identified. The media item is associated with a media class. An indication of the identified media item is provided as input to a first machine learning model. The first machine learning model is trained to predict, for a given media item, a set of encoder parameter settings that satisfy a performance criterion in view of a respective media class associated with the given media item. One or more outputs of the first machine learning model are obtained. The one or more obtained outputs include encoder data identifying one or more sets of encoder parameter settings and, for each of the sets of encoder parameter settings, an indication of a level of confidence that a respective set of encoder parameter settings satisfies the performance criterion in view of the media class associated with the identified media item.Type: GrantFiled: August 31, 2021Date of Patent: January 9, 2024Assignee: Google LLCInventors: Ching Yin Derek Pang, Kyrah Felder, Akshay Gadde, Paul Wilkins, Cheng Chen, Yao-Chung Lin
-
Publication number: 20230101806Abstract: A learning model is trained for rate-distortion behavior prediction against a corpus of a video hosting platform and used to determine optimal bitrate allocations for video data given video content complexity across the corpus of the video hosting platform. Complexity features of the video data are processed using the learning model to determine a rate-distortion cluster prediction for the video data, and transcoding parameters for transcoding the video data are selected based on that prediction. The rate-distortion clusters are modeled during the training of the learning model, such as based on rate-distortion curves of video data of the corpus of the video hosting platform and based on classifications of such video data. This approach minimizes total corpus egress and/or storage while further maintaining uniformity in the delivered quality of videos by the video hosting platform.Type: ApplicationFiled: May 19, 2020Publication date: March 30, 2023Inventors: Sam John, Balineedu Adsumilli, Akshay Gadde
-
Publication number: 20230068026Abstract: Methods and systems for encoder parameter setting optimization. A media item to be provided to one or more users of a platform is identified. The media item is associated with a media class. An indication of the identified media item is provided as input to a first machine learning model. The first machine learning model is trained to predict, for a given media item, a set of encoder parameter settings that satisfy a performance criterion in view of a respective media class associated with the given media item. One or more outputs of the first machine learning model are obtained. The one or more obtained outputs include encoder data identifying one or more sets of encoder parameter settings and, for each of the sets of encoder parameter settings, an indication of a level of confidence that a respective set of encoder parameter settings satisfies the performance criterion in view of the media class associated with the identified media item.Type: ApplicationFiled: August 31, 2021Publication date: March 2, 2023Inventors: Ching Yin Derek Pang, Kyrah Felder, Akshay Gadde, Paul Wilkins, Cheng Chen, Yao-Chung Lin
-
Patent number: 11451773Abstract: A video encoder and video decoder may determine a set of adaptive loop filters, from among a plurality of sets of adaptive loop filters, on a per-block basis. Each set of adaptive loop filters may include filters from a previous picture, filters signaled for the current picture, and/or pre-trained filter. By varying the set of adaptive loop filters on a per-block basis, the adaptive loop filters available for each block of video data may be more adapted to local statistics of the video data.Type: GrantFiled: May 30, 2019Date of Patent: September 20, 2022Assignee: QUALCOMM IncorporatedInventors: Marta Karczewicz, Akshay Gadde, Vadim Seregin, Wei-Jung Chien
-
Patent number: 11218694Abstract: Techniques are described for improving transform coding. For example, an encoded block of video data can be obtained, and a width and/or a height of the block can be determined. The width can be compared to a first threshold and/or the height can be compared to a second threshold. A horizontal transform and a vertical transform can be determined for the block based on comparing the width of the block to the first threshold and/or the height of the block to the second threshold. The horizontal transform and the vertical transform are determined without decoding a syntax element that indicates the horizontal transform and the vertical transform (e.g., the syntax element is not in an encoded video bitstream processed by a decoding device). In some cases, residual data is determined using the horizontal and vertical transforms, and a video block is determined using the residual data and a predictive block.Type: GrantFiled: September 19, 2019Date of Patent: January 4, 2022Assignee: QUALCOMM IncorporatedInventors: Vadim Seregin, Akshay Gadde, Hilmi Enes Egilmez, Marta Karczewicz, Amir Said
-
Patent number: 11140418Abstract: A video decoder can be configured to set each block-level syntax element of a plurality of block-level syntax elements to a value indicating that an adaptive loop filter is enabled for an associated component of the video data in response to determining that a slice-level syntax element is set to a value indicating that values for the plurality of block-level syntax elements are inferred.Type: GrantFiled: July 15, 2019Date of Patent: October 5, 2021Assignee: Qualcomm IncorporatedInventors: Marta Karczewicz, Nan Hu, Vadim Seregin, Akshay Gadde
-
Patent number: 10972729Abstract: Codecs that use larger blocks may have larger boundary regions that may benefit from filtering. In some embodiments, the deblocking filter determines filters and/or a number of samples of the block to be filtered based on block dimensions. For example, In one embodiment, deblocking filter parameters for the video block are determined based on at least one dimension of the size of the video block. The filter parameters include a filter to be applied or a number of pixels along a boundary with a neighboring block to which the filter is to be applied determined based on the at least one dimension.Type: GrantFiled: January 17, 2019Date of Patent: April 6, 2021Assignee: QUALCOMM IncorporatedInventors: Akshay Gadde, Li Zhang, Wei-Jung Chien, Marta Karczewicz
-
Publication number: 20200099924Abstract: Techniques are described for improving transform coding. For example, an encoded block of video data can be obtained, and a width and/or a height of the block can be determined. The width can be compared to a first threshold and/or the height can be compared to a second threshold. A horizontal transform and a vertical transform can be determined for the block based on comparing the width of the block to the first threshold and/or the height of the block to the second threshold. The horizontal transform and the vertical transform are determined without decoding a syntax element that indicates the horizontal transform and the vertical transform (e.g., the syntax element is not in an encoded video bitstream processed by a decoding device). In some cases, residual data is determined using the horizontal and vertical transforms, and a video block is determined using the residual data and a predictive block.Type: ApplicationFiled: September 19, 2019Publication date: March 26, 2020Inventors: Vadim SEREGIN, Akshay GADDE, Hilmi Enes EGILMEZ, Marta KARCZEWICZ, Amir SAID
-
Publication number: 20200029095Abstract: A video decoder can be configured to set each block-level syntax element of a plurality of block-level syntax elements to a value indicating that an adaptive loop filter is enabled for an associated component of the video data in response to determining that a slice-level syntax element is set to a value indicating that values for the plurality of block-level syntax elements are inferred.Type: ApplicationFiled: July 15, 2019Publication date: January 23, 2020Inventors: Marta Karczewicz, Nan Hu, Vadim Seregin, Akshay Gadde
-
Publication number: 20190373292Abstract: A video decoder is configured to determine, for a first block of video data, that a filter process using a first filter for a sub-block of the first block uses pixel values from neighboring sub-blocks of the sub-block of the first block; determine a number of sub-blocks in the first block of video data; and in response to the number of sub-blocks in the first block of video data being equal to one, filter the sub-block using the first filter.Type: ApplicationFiled: May 30, 2019Publication date: December 5, 2019Inventors: Akshay Gadde, Vadim Seregin, Wei-Jung Chien, Marta Karczewicz
-
Publication number: 20190373258Abstract: A video encoder and video decoder may determine a set of adaptive loop filters, from among a plurality of sets of adaptive loop filters, on a per-block basis. Each set of adaptive loop filters may include filters from a previous picture, filters signaled for the current picture, and/or pre-trained filter. By varying the set of adaptive loop filters on a per-block basis, the adaptive loop filters available for each block of video data may be more adapted to local statistics of the video data.Type: ApplicationFiled: May 30, 2019Publication date: December 5, 2019Inventors: Marta Karczewicz, Akshay Gadde, Vadim Seregin, Wei-Jung Chien
-
Publication number: 20190320172Abstract: A video decoder configured to determine filter information for a region of a picture of video data; for a largest coding unit (LCU) of the region, determine a size for a current unit based on a location of the LCU within the picture, wherein the size for the current unit is different than a size of the LCU; and based on the determined filter information, determine a filter for the current unit of the region; and filter the current unit with the determined filter.Type: ApplicationFiled: April 11, 2019Publication date: October 17, 2019Inventors: Li Zhang, Cheng-Teh Hsieh, Kai Zhang, Akshay Gadde, Wei-Jung Chien, Marta Karczewicz
-
Publication number: 20190306502Abstract: Methods and systems for improved adaptive loop filters (ALFs) used in post-processing stage of in-loop coding or the prediction stage of video coding. To account for various shortcomings, techniques to improve coding gains and visual quality of ALFs are discussed. First, refinement of ALF coefficients for each block is allowed wherein different units (used for class calculation, e.g., 2×2 sub-blocks in GALF) located in different blocks with the same class index may have different filters. Second, ALF filters can be modified or weakened to without signaling ALF filter coefficients.Type: ApplicationFiled: April 1, 2019Publication date: October 3, 2019Inventors: Akshay GADDE, Li Zhang, Marta Karczewicz, Wei-Jung Chein
-
Publication number: 20190230353Abstract: Codecs that use larger blocks may have larger boundary regions that may benefit from filtering. In some embodiments, the deblocking filter determines filters and/or a number of samples of the block to be filtered based on block dimensions. For example, In one embodiment, deblocking filter parameters for the video block are determined based on at least one dimension of the size of the video block. The filter parameters include a filter to be applied or a number of pixels along a boundary with a neighboring block to which the filter is to be applied determined based on the at least one dimension.Type: ApplicationFiled: January 17, 2019Publication date: July 25, 2019Inventors: Akshay GADDE, Li Zhang, Wei-Jung Chien, Marta Karczewicz
-
Patent number: 9510787Abstract: A method reconstructs a signal by sampling the signal using a sampling procedure to obtain an input signal. A consistent set is determined from the input signal including the first elements such that applying the sampling procedure to the first elements results in the input signal. According to the type of the signal, a guiding set is determined including second elements disjoint from the first elements. A reconstruction set including third elements is generated so that the third elements minimize a sum of a first similarity measure of the third elements to the second elements and a second similarity measure of the third elements to the first elements. A transformed signal that minimizes a function on the reconstruction set is determined. A reconstructed signal is rendered so that a third similarity measure of the reconstructed signal to the transformed signal is smaller than a tolerance.Type: GrantFiled: December 11, 2014Date of Patent: December 6, 2016Assignee: Mitsubishi Electric Research Laboratories, Inc.Inventors: Andrei Kniazev, Akshay Gadde, Dong Tian, Hassan Mansour, Richard C Waters
-
Publication number: 20160173736Abstract: A method reconstructs a signal by sampling the signal using a sampling procedure to obtain an input signal. A consistent set is determined from the input signal including the first elements such that applying the sampling procedure to the first elements results in the input signal. According to the type of the signal, a guiding set is determined including second elements disjoint from the first elements. A reconstruction set including third elements is generated so that the third elements minimize a sum of a first similarity measure of the third elements to the second elements and a second similarity measure of the third elements to the first elements. A transformed signal that minimizes a function on the reconstruction set is determined. A reconstructed signal is rendered so that a third similarity measure of the reconstructed signal to the transformed signal is smaller than a tolerance.Type: ApplicationFiled: December 11, 2014Publication date: June 16, 2016Inventors: Andrei Kniazev, Akshay Gadde, Dong Tian, Hassan Mansour, Richard C. Waters