Patents by Inventor Chaitanya Satish Ghone
Chaitanya Satish Ghone 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: 20230283808Abstract: A method is provided that includes receiving pictures of a video sequence in a video encoder, and encoding the pictures to generate a compressed video bit stream that is transmitted to a video decoder in real-time, wherein encoding the pictures includes selecting a picture to be encoded as a delayed duplicate intra-predicted picture (DDI), wherein the picture would otherwise be encoded as an inter-predicted picture (P-picture), encoding the picture as an intra-predicted picture (I-picture) to generate the DDI, wherein the I-picture is reconstructed and stored for use as a reference picture for a decoder refresh picture, transmitting the DDI to the video decoder in non-real time, selecting a subsequent picture to be encoded as the decoder refresh picture, and encoding the subsequent picture in the compressed bit stream as the decoder refresh picture, wherein the subsequent P-picture is encoded as a P-picture predicted using the reference picture.Type: ApplicationFiled: May 15, 2023Publication date: September 7, 2023Inventors: Ranga Ramanujam Srinivasan, Chaitanya Satish Ghone, Mihir Narendra Mody, Minhua Zhou
-
Patent number: 11653031Abstract: A method is provided that includes receiving pictures of a video sequence in a video encoder, and encoding the pictures to generate a compressed video bit stream that is transmitted to a video decoder in real-time, wherein encoding the pictures includes selecting a picture to be encoded as a delayed duplicate intra-predicted picture (DDI), wherein the picture would otherwise be encoded as an inter-predicted picture (P-picture), encoding the picture as an intra-predicted picture (I-picture) to generate the DDI, wherein the I-picture is reconstructed and stored for use as a reference picture for a decoder refresh picture, transmitting the DDI to the video decoder in non-real time, selecting a subsequent picture to be encoded as the decoder refresh picture, and encoding the subsequent picture in the compressed bit stream as the decoder refresh picture, wherein the subsequent P-picture is encoded as a P-picture predicted using the reference picture.Type: GrantFiled: November 10, 2020Date of Patent: May 16, 2023Assignee: Texas Instruments IncorporatedInventors: Ranga Ramanujam Srinivasan, Chaitanya Satish Ghone, Mihir Narendra Mody, Minhua Zhou
-
Patent number: 11443505Abstract: A method for analyzing images to generate a plurality of output features includes receiving input features of the image and performing Fourier transforms on each input feature. Kernels having coefficients of a plurality of trained features are received and on-the-fly Fourier transforms (OTF-FTs) are performed on the coefficients in the kernels. The output of each Fourier transform and each OTF-FT are multiplied together to generate a plurality of products and each of the products are added to produce one sum for each output feature. Two-dimensional inverse Fourier transforms are performed on each sum.Type: GrantFiled: June 11, 2020Date of Patent: September 13, 2022Assignee: TEXAS INSTRUMENTS INCORPORATEDInventors: Mihir Narendra Mody, Manu Mathew, Chaitanya Satish Ghone
-
Publication number: 20210058645Abstract: A method is provided that includes receiving pictures of a video sequence in a video encoder, and encoding the pictures to generate a compressed video bit stream that is transmitted to a video decoder in real-time, wherein encoding the pictures includes selecting a picture to be encoded as a delayed duplicate intra-predicted picture (DDI), wherein the picture would otherwise be encoded as an inter-predicted picture (P-picture), encoding the picture as an intra-predicted picture (I-picture) to generate the DDI, wherein the I-picture is reconstructed and stored for use as a reference picture for a decoder refresh picture, transmitting the DDI to the video decoder in non-real time, selecting a subsequent picture to be encoded as the decoder refresh picture, and encoding the subsequent picture in the compressed bit stream as the decoder refresh picture, wherein the subsequent P-picture is encoded as a P-picture predicted using the reference picture.Type: ApplicationFiled: November 10, 2020Publication date: February 25, 2021Inventors: Ranga Ramanujam Srinivasan, Chaitanya Satish Ghone, Mihir Narendra Mody, Minhua Zhou
-
Patent number: 10869064Abstract: A method is provided that includes receiving pictures of a video sequence in a video encoder, and encoding the pictures to generate a compressed video bit stream that is transmitted to a video decoder in real-time, wherein encoding the pictures includes selecting a picture to be encoded as a delayed duplicate intra-predicted picture (DDI), wherein the picture would otherwise be encoded as an inter-predicted picture (P-picture), encoding the picture as an intra-predicted picture (I-picture) to generate the DDI, wherein the I-picture is reconstructed and stored for use as a reference picture for a decoder refresh picture, transmitting the DDI to the video decoder in non-real time, selecting a subsequent picture to be encoded as the decoder refresh picture, and encoding the subsequent picture in the compressed bit stream as the decoder refresh picture, wherein the subsequent P-picture is encoded as a P-picture predicted using the reference picture.Type: GrantFiled: February 26, 2018Date of Patent: December 15, 2020Assignee: TEXAS INSTRUMENTS INCORPORATEDInventors: Ranga Ramanujam Srinivasan, Chaitanya Satish Ghone, Mihir Narendra Mody, Minhua Zhou
-
Publication number: 20200302217Abstract: A method for analyzing images to generate a plurality of output features includes receiving input features of the image and performing Fourier transforms on each input feature. Kernels having coefficients of a plurality of trained features are received and on-the-fly Fourier transforms (OTF-FTs) are performed on the coefficients in the kernels. The output of each Fourier transform and each OTF-FT are multiplied together to generate a plurality of products and each of the products are added to produce one sum for each output feature. Two-dimensional inverse Fourier transforms are performed on each sum.Type: ApplicationFiled: June 11, 2020Publication date: September 24, 2020Inventors: Mihir Narendra Mody, Manu Mathew, Chaitanya Satish Ghone
-
Patent number: 10713522Abstract: A method for analyzing images to generate a plurality of output features includes receiving input features of the image and performing Fourier transforms on each input feature. Kernels having coefficients of a plurality of trained features are received and on-the-fly Fourier transforms (OTF-FTs) are performed on the coefficients in the kernels. The output of each Fourier transform and each OTF-FT are multiplied together to generate a plurality of products and each of the products are added to produce one sum for each output feature. Two-dimensional inverse Fourier transforms are performed on each sum.Type: GrantFiled: May 1, 2019Date of Patent: July 14, 2020Assignee: TEXAS INSTRUMENTS INCORPORATEDInventors: Mihir Narendra Mody, Manu Mathew, Chaitanya Satish Ghone
-
Patent number: 10397591Abstract: A control processor for a video encode-decode engine is provided that includes an instruction pipeline. The instruction pipeline includes an instruction fetch stage coupled to an instruction memory to fetch instructions, an instruction decoding stage coupled to the instruction fetch stage to receive the fetched instructions, and an execution stage coupled to the instruction decoding stage to receive and execute decoded instructions. The instruction decoding stage and the instruction execution stage are configured to decode and execute a set of instructions in an instruction set of the control processor that are designed specifically for accelerating video sequence encoding and encoded video bit stream decoding.Type: GrantFiled: April 11, 2015Date of Patent: August 27, 2019Assignee: TEXAS INSTRUMENTS INCORPORATEDInventors: Dipan Kumar Mandal, Mihir Narendra Mody, Mahesh Madhukar Mehendale, Chaitanya Satish Ghone, Piyali Goswami, Naresh Kumar Yadav, Hetul Sanghvi, Niraj Nandan
-
Publication number: 20190258891Abstract: A method for analyzing images to generate a plurality of output features includes receiving input features of the image and performing Fourier transforms on each input feature. Kernels having coefficients of a plurality of trained features are received and on-the-fly Fourier transforms (OTF-FTs) are performed on the coefficients in the kernels. The output of each Fourier transform and each OTF-FT are multiplied together to generate a plurality of products and each of the products are added to produce one sum for each output feature. Two-dimensional inverse Fourier transforms are performed on each sum.Type: ApplicationFiled: May 1, 2019Publication date: August 22, 2019Inventors: Mihir Narendra Mody, Manu Mathew, Chaitanya Satish Ghone
-
Patent number: 10325173Abstract: A method for analyzing images to generate a plurality of output features includes receiving input features of the image and performing Fourier transforms on each input feature. Kernels having coefficients of a plurality of trained features are received and on-the-fly Fourier transforms (OTF-FTs) are performed on the coefficients in the kernels. The output of each Fourier transform and each OTF-FT are multiplied together to generate a plurality of products and each of the products are added to produce one sum for each output feature. Two-dimensional inverse Fourier transforms are performed on each sum.Type: GrantFiled: August 22, 2018Date of Patent: June 18, 2019Assignee: TEXAS INSTRUMENTS INCORPORATEDInventors: Mihir Narendra Mody, Manu Mathew, Chaitanya Satish Ghone
-
Publication number: 20180357513Abstract: A method for analyzing images to generate a plurality of output features includes receiving input features of the image and performing Fourier transforms on each input feature. Kernels having coefficients of a plurality of trained features are received and on-the-fly Fourier transforms (OTF-FTs) are performed on the coefficients in the kernels. The output of each Fourier transform and each OTF-FT are multiplied together to generate a plurality of products and each of the products are added to produce one sum for each output feature. Two-dimensional inverse Fourier transforms are performed on each sum.Type: ApplicationFiled: August 22, 2018Publication date: December 13, 2018Inventors: Mihir Narendra Mody, Manu Mathew, Chaitanya Satish Ghone
-
Patent number: 10083374Abstract: A method for analyzing images to generate a plurality of output features includes receiving input features of the image and performing Fourier transforms on each input feature. Kernels having coefficients of a plurality of trained features are received and on-the-fly Fourier transforms (OTF-FTs) are performed on the coefficients in the kernels. The output of each Fourier transform and each OTF-FT are multiplied together to generate a plurality of products and each of the products are added to produce one sum for each output feature. Two-dimensional inverse Fourier transforms are performed on each sum.Type: GrantFiled: December 12, 2016Date of Patent: September 25, 2018Assignee: TEXAS INSTRUMENTS INCORPORATEDInventors: Mihir Narendra Mody, Manu Mathew, Chaitanya Satish Ghone
-
Publication number: 20180184130Abstract: A method is provided that includes receiving pictures of a video sequence in a video encoder, and encoding the pictures to generate a compressed video bit stream that is transmitted to a video decoder in real-time, wherein encoding the pictures includes selecting a picture to be encoded as a delayed duplicate intra-predicted picture (DDI), wherein the picture would otherwise be encoded as an inter-predicted picture (P-picture), encoding the picture as an intra-predicted picture (I-picture) to generate the DDI, wherein the I-picture is reconstructed and stored for use as a reference picture for a decoder refresh picture, transmitting the DDI to the video decoder in non-real time, selecting a subsequent picture to be encoded as the decoder refresh picture, and encoding the subsequent picture in the compressed bit stream as the decoder refresh picture, wherein the subsequent P-picture is encoded as a P-picture predicted using the reference picture.Type: ApplicationFiled: February 26, 2018Publication date: June 28, 2018Inventors: Ranga Ramanujam Srinivasan, Chaitanya Satish Ghone, Mihir Narendra Mody, Minhua Zhou
-
Publication number: 20180165543Abstract: A method for analyzing images to generate a plurality of output features includes receiving input features of the image and performing Fourier transforms on each input feature. Kernels having coefficients of a plurality of trained features are received and on-the-fly Fourier transforms (OTF-FTs) are performed on the coefficients in the kernels. The output of each Fourier transform and each OTF-FT are multiplied together to generate a plurality of products and each of the products are added to produce one sum for each output feature. Two-dimensional inverse Fourier transforms are performed on each sum.Type: ApplicationFiled: December 12, 2016Publication date: June 14, 2018Inventors: Mihir Narendra Mody, Manu Mathew, Chaitanya Satish Ghone
-
Patent number: 9906815Abstract: A method is provided that includes receiving pictures of a video sequence in a video encoder, and encoding the pictures to generate a compressed video bit stream that is transmitted to a video decoder in real-time, wherein encoding the pictures includes selecting a picture to be encoded as a delayed duplicate intra-predicted picture (DDI), wherein the picture would otherwise be encoded as an inter-predicted picture (P-picture), encoding the picture as an intra-predicted picture (I-picture) to generate the DDI, wherein the I-picture is reconstructed and stored for use as a reference picture for a decoder refresh picture, transmitting the DDI to the video decoder in non-real time, selecting a subsequent picture to be encoded as the decoder refresh picture, and encoding the subsequent picture in the compressed bit stream as the decoder refresh picture, wherein the subsequent P-picture is encoded as a P-picture predicted using the reference picture.Type: GrantFiled: November 7, 2012Date of Patent: February 27, 2018Assignee: TEXAS INSTRUMENTS INCORPORATEDInventors: Ranga Ramanujam Srinivasan, Chaitanya Satish Ghone, Mihir Narendra Mody, Minhua Zhou
-
Publication number: 20160165230Abstract: Checksum computation for video coding is provided that breaks the dependency between the color components of a picture in the prior art. More specifically, rather than computing a single checksum for a picture as in the prior art, a separate checksum is computed for each color component. Computing a separate checksum for each color component enables parallel computation of the component checksums. Methods are provided for computing three separate checksums after a picture is decoded. Methods are also provided for computing three separate checksums on a largest coding unit basis, thus allowing the checksums for a picture to be computed as the picture is being decoded.Type: ApplicationFiled: April 16, 2013Publication date: June 9, 2016Applicant: Texas Instruments IncorporatedInventors: Ranga Ramanujam Srinivasan, Mihir Narendra Mody, Chaitanya Satish Ghone
-
Publication number: 20150296212Abstract: A control processor for a video encode-decode engine is provided that includes an instruction pipeline. The instruction pipeline includes an instruction fetch stage coupled to an instruction memory to fetch instructions, an instruction decoding stage coupled to the instruction fetch stage to receive the fetched instructions, and an execution stage coupled to the instruction decoding stage to receive and execute decoded instructions. The instruction decoding stage and the instruction execution stage are configured to decode and execute a set of instructions in an instruction set of the control processor that are designed specifically for accelerating video sequence encoding and encoded video bit stream decoding.Type: ApplicationFiled: April 11, 2015Publication date: October 15, 2015Inventors: Dipan Kumar Mandal, Mihir Narendra Mody, Mahesh Madhukar Mehendale, Chaitanya Satish Ghone, Piyali Goswami, Naresh Kumar Yadav, Hetul Sanghvi, Niraj Nandan
-
Publication number: 20130272429Abstract: Checksum computation for video coding is provided that breaks the dependency between the color components of a picture in the prior art. More specifically, rather than computing a single checksum for a picture as in the prior art, a separate checksum is computed for each color component. Computing a separate checksum for each color component enables parallel computation of the component checksums. Methods are provided for computing three separate checksums after a picture is decoded. Methods are also provided for computing three separate checksums on a largest coding unit basis, thus allowing the checksums for a picture to be computed as the picture is being decoded.Type: ApplicationFiled: April 16, 2013Publication date: October 17, 2013Applicant: Texas Instruments IncorporatedInventors: Ranga Ramanujam Srinivasan, Mihir Narendra Mody, Chaitanya Satish Ghone