# US 12,762,841 B2 · Method and system for offloading decoding of encoded data from a processing system > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,762,841 B2 - Document ID: US12762841B2 - Status: Granted - Kind: B2 - Application number: 17/970,309 - Filed: 2022-10-20 - Published: 2024-04-25 - Granted: 2026-10-13 - Earliest claimed: 2022-10-20 - Assignees: Dell Products L.P. - Applicants: Dell Products L.P. - Inventors: Ian Roche; Philip Hummel; Dharmesh M. Patel - Web page: https://patentgrep.com/patent/US12762841B2 ## CPC - G06V 10/766: using regression, e.g. by projecting features on hyperplanes - G06T 1/20: Processor architectures; Processor configuration, e.g. pipelining - G06V 10/12: Details of acquisition arrangements; Constructional details thereof - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL ## Abstract A system includes a processor, a processing system operatively connected to the processor, comprising a decoding-dedicated hardware component, an operating system operating on the processor and not on the processing system, wherein the processing system is programmed to obtain encoded data from a local data source, perform, using the decoding-dedicated hardware component, a decoding of the encoded data to obtain decoded data, perform a computer vision (CV) inferencing on the decoded data to obtain inferencing data, provide the inferencing data to the processor, and perform, by the processor, a remediation action based on the inferencing data. ## Claims ### Claim 1 A system, comprising: a processor; a processing system operatively connected to the processor, comprising a decoding-dedicated hardware component; an operating system operating on the processor and not on the processing system; wherein the processing system is programmed to: obtain, from an enhanced networking interface, encoded data associated with a video stream from a local data source, wherein the enhanced networking interface performs initial processing on the encoded data, wherein the initial processing comprises converting the video stream from an analog format to a digital format; wherein the encoded data is obtained from the enhanced networking interface via remote direct memory access (RDMA), and wherein the enhanced networking interface obtains the video stream from the local data source via the RDMA; perform, using the decoding-dedicated hardware component, a decoding of the encoded data to obtain decoded data; perform a computer vision (CV) inferencing on the decoded data to obtain inferencing data, wherein the inferencing is performed based on a federated CV regression model obtained from a CV manager operatively connected to the processing system via a network, and wherein the federated CV regression model is generated at least partially using a local CV regression generated by the processing system, wherein the local CV regression comprises a first set of weights, and wherein the federated CV regression comprises a second set of weights based on: the first set of weights, a third set of weights of a second local CV regression model generated by a second processing system, and a fourth set of weights associated with an initial federated CV regression model, wherein the initial federated CV regression model is generated based on a training of historical data provided to the processing system, and wherein the federated CV regression model is generated by applying an average of: the first set of weights, the second set of weights, and the fourth set of weights associated with an initial federated CV regression model; provide the inferencing data to the processor; generate metadata corresponding to hardware of the local data source; and perform, by the processor, a remediation action based on the inferencing data and based on the metadata, wherein the remediation action comprises increasing a number of video cameras and a number of graphics processing units (GPUs) used by the local data source for CV inferencing. ### Claim 2 (depends on 1) The system of claim 1, wherein the processing system further comprises a graphics processing unit (GPU). ### Claim 3 (depends on 1) The system of claim 1, wherein the encoded data is video stream data. ### Claim 4 (depends on 3) The system of claim 3, wherein the video stream data is generated by a video camera, and wherein the encoded data is encoded using a video management system (VMS) application. ### Claim 5 (depends on 1) The system of claim 1, wherein performing the decoding comprises formatting the encoded data from a first format to a second format readable by the processing system. ### Claim 6 (depends on 1) The system of claim 1, wherein the decoding-dedicated hardware component is a field programmable gate array (FPGA). ### Claim 7 A non-transitory computer readable medium comprising computer readable program code, which when executed by a computer processor enables the computer processor to perform a method for managing information handling systems, the method comprising: obtaining, by a processing system and from an enhanced networking interface, encoded data associated with a video stream from a local data source, wherein the enhanced networking interface performs initial processing on the encoded data, wherein the initial processing comprises converting the video stream from an analog format to a digital format; wherein the encoded data is obtained from the enhanced networking interface via remote direct memory access (RDMA), and wherein the enhanced networking interface obtains the video stream from the local data source via the RDMA; performing, using decoding-dedicated hardware component of the processing system, a decoding of the encoded data to obtain decoded data; performing a computer vision (CV) inferencing on the decoded data to obtain inferencing data, wherein the inferencing is performed based on a federated CV regression model obtained from a CV manager operatively connected to the processing system via a network, and wherein the federated CV regression model is generated at least partially using a local CV regression generated by the processing system, wherein the local CV regression comprises a first set of weights, and wherein the federated CV regression comprises a second set of weights based on: the first set of weights, a third set of weights of a second local CV regression model generated by a second processing system, and a fourth set of weights associated with an initial federated CV regression model, wherein the initial federated CV regression model is generated based on a training of historical data provided to the processing system, and wherein the federated CV regression model is generated by applying an average of: the first set of weights, the second set of weights, and the fourth set of weights; providing the inferencing data to a processor; generating metadata corresponding to hardware of the local data source; and performing, by the processor, a remediation action based on the inferencing data and based on the metadata, wherein an operating system operates on the processor and not on the processing system, and wherein the remediation action comprises increasing a number of video cameras and a number of graphics processing units (GPUs) used by the local data source for CV inferencing. ### Claim 8 (depends on 7) The non-transitory computer readable medium of claim 7, wherein the processing system further comprises a GPU. ### Claim 9 (depends on 7) The non-transitory computer readable medium of claim 7, wherein the encoded data is video stream data. ### Claim 10 (depends on 9) The non-transitory computer readable medium of claim 9, wherein the video stream data is generated by a video camera, and wherein the encoded data is encoded using a video management system (VMS) application. ### Claim 11 (depends on 7) The non-transitory computer readable medium of claim 7, wherein performing the decoding comprises formatting the encoded data from a first format to a second format readable by the processing system. ### Claim 12 (depends on 7) The non-transitory computer readable medium of claim 7, wherein the decoding-dedicated hardware component is a field programmable gate array (FPGA). ### Claim 13 A method for managing hardware resources, comprising: obtaining, by a processing system and from an enhanced networking interface, encoded data associated with a video stream from a local data source, wherein the enhanced networking interface performs initial processing on the encoded data, wherein the initial processing comprises converting the video stream from an analog format to a digital format; wherein the encoded data is obtained from the enhanced networking interface via remote direct memory access (RDMA), and wherein the enhanced networking interface obtains the video stream from the local data source via the RDMA; performing, using decoding-dedicated hardware component of the processing system, a decoding of the encoded data to obtain decoded data; performing a computer vision (CV) inferencing on the decoded data to obtain inferencing data, wherein the inferencing is performed based on a federated CV regression model obtained from a CV manager operatively connected to the processing system via a network, and wherein the federated CV regression model is generated at least partially using a local CV regression generated by the processing system, wherein the local CV regression comprises a first set of weights, and wherein the federated CV regression comprises a second set of weights based on: the first set of weights, a third set of weights of a second local CV regression model generated by a second processing system, and a fourth set of weights associated with an initial federated CV regression model, wherein the initial federated CV regression model is generated based on a training of historical data provided to the processing system, and wherein the federated CV regression model is generated by applying an average of: the first set of weights, the second set of weights, and the fourth set of weights associated with an initial federated CV regression model; providing the inferencing data to a processor; generating metadata corresponding to hardware of the local data source; and performing, by the processor, a remediation action based on the inferencing data and based on the metadata, wherein an operating system operates on the processor and not on the processing system, and wherein the remediation action comprises increasing a number of video cameras and a number of graphics processing units (GPUs) used by the local data source for CV inferencing. ### Claim 14 (depends on 13) The method of claim 13, wherein the processing system further comprises a graphics processing unit (GPU). ### Claim 15 (depends on 13) The method of claim 13, wherein the encoded data is video stream data. ### Claim 16 (depends on 13) The method of claim 13, wherein the video stream data is generated by a video camera, and wherein the encoded data is encoded using a video management system (VMS) application. ### Claim 17 (depends on 13) The method of claim 13, wherein performing the decoding comprises formatting the encoded data from a first format to a second format readable by the processing system. ### Claim 18 (depends on 13) The method of claim 13, wherein the decoding-dedicated hardware component is a field programmable gate array (FPGA).