Granted Patent B2

US 12,647,584 · App. 18/249,043 · Granted Jun 2, 2026

Inventors: Guido Meardi (London, GB); Laurence Charles Venner (London, GB)

Assignee: V-NOVA INTERNATIONAL LIMITED

H04N19/33G06V10/95G06V20/44H04N7/181

Abstract

Claims (68)

1 . A method of analysing a plurality of video camera feeds, the method comprising:

encoding, at a first location, the plurality of video camera feeds using a layer-based encoding, including generating encoded data streams for each of a plurality of layers within the layer-based encoding, wherein different layers in the plurality of layers correspond to different levels of quality, higher layers representing higher levels of quality;

transmitting, to a second location remote from the first location, encoded data streams for one or more lowest layers for the plurality of video camera feeds;

decoding, at the second location, the encoded data streams to generate a set of reconstructions of the plurality of video camera feeds at a first level of quality;

applying one or more video analysis functions to the set of reconstructions to identify one or more video camera feeds for further analysis;

sending, to the first location for the identified one or more video camera feeds for further analysis, a request for further encoded data streams for one or more layers above the one or more lowest layers;

responsive to the request, transmitting, to the second location, the further encoded data streams for one or more layers above the one or more lowest layers;

decoding, at the second location, the further encoded data streams to generate a set of reconstructions for the identified one or more video camera feeds at a second level of quality; and

applying one or more video analysis functions to the set of reconstructions at the second level of quality,

wherein the layer-based encoding comprises a base layer at a lowest level of quality and one or more residual layers above the base layer at higher levels of quality, a given residual layer in the one or more residual layers being determined based on a comparison of data derived from an input signal at a level of quality for the layer and data derived from a lower layer, and

wherein the one or more residual layers are generated without inter-block dependencies such that each coding unit within each residual layer can be decoded independently of other coding units within the same residual layer, and wherein each residual layer is decoded independently of other residual layers and the base layer.

2 . The method of claim 1 , further comprising:

determining a region of interest in at least one of reconstructions of the plurality of input signals at the one or more lowest layers; and

requesting further encoded data relating to the region of interest.

3 . The method of claim 2 , further comprising:

generating a set of reconstructions for the region of interest at one or more layers above the one or more lowest layers using the further encoded data.

4 . The method of claim 2 , further comprising:

determining differential encoded data required to generate a reconstruction at a highest level of quality, the differential encoded data excluding the received encoded data and the received further encoded data; and

requesting the differential encoded data from the first location.

5 . The method of claim 1 , further comprising:

performing object detection within at least one of reconstructions of the plurality of input signals at the one or more lowest layers.

6 . The method of claim 1 , further comprising:

performing object detection within at least one of reconstructions of the plurality of input signals at one or more layers above the one or more lowest layers.

7 . The method of claim 1 , the method comprising one or more of:

identifying one or more sets of frames for one or more of the plurality of input signals for further analysis, and

identifying one or more input sources for further analysis;

wherein said identifying is performed at one or more of the first, encoding, location and the second, analysis location.

8 . The method of claim 7 , further comprising:

transmitting encoded data for layers above the one or more lowest layers for the identified frames or the identified input sources.

9 . The method of claim 1 , further comprising:

transmitting encoded data to a remote server for signal analysis; and

transmitting encoded data to a plurality of users.

10 . The method of claim 9 , wherein transmitting encoded data to a plurality of users comprises transmitting encoded data to a distribution centre for distribution to the plurality of users.

11 . The method of claim 10 , further comprising:

analysing a set of reconstructions at the remote server to determine a set of layers of encoded data for one or more input sources to be transmitted to the distribution centre, wherein the plurality of input signals or the plurality of video input feeds are associated with a plurality of video cameras to capture a live event, wherein the remote server determines which of the plurality of video cameras to select for transmission to the plurality of users, and the levels of quality for the encoded data to transmit to the plurality of users for the selected video cameras.

12 . The method of claim 1 , wherein the encoded data transmitted from the first, encoding location, comprises encoded data for a subset of encoded signal components, wherein the encoded signal components comprise different colour components of the video signal and the subset of encoded components comprise at least a lightness or luma component.

13 . The method of claim 1 , further comprising:

receiving at the second, remote, location, encoded data for one or more layers above the one or more lowest layers for archiving.

14 . The method of claim 1 , further comprising:

decoding the received encoded data to obtain a plurality of reconstructions of the plurality of input signals at the one or more lowest layers;

upscaling the plurality of reconstructions;

performing analysis on the upscaled reconstructions,

wherein the upscaling uses a neural network architecture trained on a training set comprise versions on an image at two spatial resolutions, so as to learn parameters values to map a reconstruction at a lower resolution to a reconstruction at a higher resolution, and

wherein the upscaling is performed responsive to encoded data for one or more layers at the higher resolution and above being unavailable.

15 . The method of claim 1 , wherein the multi-layer encoding comprises an encoding based on one or more of SMPTE VC-6 2117 and ISO/IEC MPEG-5 Part 2 LCEVC.

16 . A method of remote signal analysis comprising:

receiving, at a computing device over at least one network, encoded data for a plurality of encoded signal streams corresponding to a plurality of input signals,

wherein the encoding of the plurality of input signals is performed at an encoding location remote from the computing device,

wherein the encoded signal streams are encoded using a layer-based encoding,

wherein different layers in the layer-based encoding correspond to different levels of quality, higher layers representing higher levels of quality, and

wherein the received encoded data comprises encoded data associated with one or more lowest layers within the layer-based encoding;

decoding the received encoded data to obtain a plurality of reconstructions of the plurality of input signals at the one or more lowest layers;

processing at least one of reconstructions of the plurality of input signals to determine portions of the encoded signal streams for further processing; and

requesting, from the encoding location, encoded data for one or more layers above the one or more lowest layers that are associated with the determined portions of the encoded signal streams,

wherein the layer-based encoding comprises a base layer at a lowest level of quality and one or more residual layers above the base layer at higher levels of quality, a given residual layer in the one or more residual layers being determined based on a comparison of data derived from an input signal at a level of quality for the layer and data derived from a lower layer, and

wherein the one or more residual layers are generated without inter-block dependencies such that each coding unit within each residual layer can be decoded independently of other coding units within the same residual layer, and wherein each residual layer is decoded independently of other residual layers and the base layer.

17 . The method of claim 16 , wherein different levels of quality correspond to different spatial resolutions, higher layers representing higher spatial resolutions.

18 . The method of claim 17 , wherein each level of quality has a different encoding bit rate.

19 . An encoding method comprising:

generating, at a first location, encoded data for a plurality of encoded signal streams corresponding to a plurality of input signals,

wherein the encoded signal streams are encoded using a layer-based encoding,

wherein different layers in the layer-based encoding correspond to different levels of quality, higher layers representing higher levels of quality, and

wherein the received encoded data comprises encoded data associated with one or more lowest layers within the layer-based encoding;

transmitting the encoded data to a second location remote from the first location for decoding;

receiving a request for further encoded data for one or more layers above the one or more lowest layers, the request relating to one or more of the plurality of input signals; and

transmitting the further encoded data to the second location,

wherein the layer-based encoding comprises a base layer at a lowest level of quality and one or more residual layers above the base layer at higher levels of quality, a given residual layer in the one or more residual layers being determined based on a comparison of data derived from an input signal at a level of quality for the layer and data derived from a lower layer, and

wherein the one or more residual layers are generated without inter-block dependencies such that each coding unit within each residual layer can be decoded independently of other coding units within the same residual layer, and wherein each residual layer is decoded independently of other residual layers and the base layer.

Assignments (2)

Priority Claims (1)

Continuity (1)

References Cited (13)