Methods and devices for decoding at least part of a data stream, computer program and associated data streams
A part of a data stream includes a plurality of data units respectively associated with a plurality of images and together presenting different images of the plurality of images. The data stream part further includes differential encoding data differentially encoding an artificial neural decoding network relatively to a reference artificial neural decoding network. A method for decoding this data stream part includes the following steps: determining the artificial neural decoding network by decoding the differential coding data; and decoding at least one data unit of the plurality of data units by the determined artificial neural decoding network. Another decoding method, decoding devices, a computer program and associated data streams are also described.
1 . A decoding method for decoding a part of a data stream, said data stream part comprising a plurality of data units respectively associated with a plurality of images and representative together of the different images of the plurality of images, said data stream part further comprising differential encoding data differentially encoding a decoding artificial neural network relatively to a reference decoding artificial neural network, the decoding method being performed by an electronic device and comprising:
determining the decoding artificial neural network by decoding said differential encoding data; and
decoding at least one of the plurality of data units by means of the determined decoding artificial neural network.
2 . The decoding method according to claim 1 , wherein data characteristic of the reference decoding artificial neural network are stored in a memory of the electronic device.
3 . The decoding method according to claim 2 , wherein the data stream comprises an identifier of the reference decoding artificial neural network.
4 . The decoding method according to claim 1 , wherein the data stream comprises descriptive data of the reference decoding artificial neural network.
5 . The decoding method according to claim 4 , wherein said descriptive data are included in said data stream part.
6 . The decoding method according to claim 1 , wherein said decoding said data unit produces a two-dimensional representation of the image associated with said data unit by decoding said data unit and at least another said data unit included in said data stream part.
7 . The decoding method according to claim 1 , wherein the reference decoding artificial neural network is defined by a plurality of first weights, wherein the determined decoding artificial neural network is defined by a plurality of second weights and wherein said differential encoding data comprise data each representing a difference between a given one of the plurality of first weights and a corresponding one of the plurality of second weights.
8 . The decoding method according to claim 1 , wherein data indicative of the reference artificial neural network are contained in a set of parameters relating to said part of the data stream and wherein the differential encoding data are contained in a set of parameters relating to the image associated with said data unit.
9 . The decoding method according to claim 1 , wherein the data stream comprises at least one flag signaling presence, in part at least of the data stream, of at least another flag signaling presence of data indicative of another decoding artificial neural network, distinct from the determined decoding artificial neural network.
10 . The decoding method according to claim 1 , wherein the data stream comprises at least one flag signaling presence of said differential encoding data in said data stream part.
11 . A decoding device comprising:
at least one processor; and
at least one non-transitory computer readable medium comprising instructions stored thereon which when executed by the at least one processor configure the decoding device to decode a part of a data stream, said part of the data stream comprising a plurality of data units respectively associated with a plurality of images and representative together of the different images of the plurality of images, said part of the data stream further comprising differential encoding data differentially encoding a decoding artificial neural network relatively to a reference decoding artificial neural network, the decoding:
determining the decoding artificial neural network by decoding said differential encoding data; and
decoding at least one of the plurality of data units by means of the decoding artificial neural network determined by the determination module.
12 . The decoding device according to claim 11 , wherein the instructions configure the decoding device to store a data characteristic of the reference decoding artificial neural network.
13 . The decoding device according to claim 11 , wherein the instructions configure the decoding device to process data descriptive of the reference decoding artificial neural network included in the data stream.
14 . The decoding device according to claim 11 , wherein the instructions configure the decoding device to produce a two-dimensional representation of one of the plurality of images associated with said data unit by decoding said data unit and at least another one of the plurality of data units included in said data stream part.
15 . A non-transitory computer-readable medium on which is stored a computer program comprising instructions executable by a processor and adapted to implement the decoding method according to claim 1 when the instructions are executed by the processor.
16 . A method implemented by an encoding device and comprising:
generating at least one part of a data stream comprising a plurality of data units respectively associated with a plurality of images and representative together of the different images of the plurality of images, said at least one part further comprising differential encoding data differentially encoding a decoding artificial neural network relatively to a reference decoding artificial neural network, wherein at least one data unit of the plurality of data units can be decoded using the decoding artificial neural network encoded by said differential encoding data, and
storing and/or transmitting said data stream.
17 . The decoding method according to claim 1 , wherein the differential encoding data includes:
data respectively representative of differences between each weight of the decoding artificial neural network and a corresponding weight of the reference decoding artificial neural network, or
data representative of a subset of weights of the reference decoding artificial neural network, or
data representative of additional layers of artificial neural network to be inserted within the structure of the reference decoding artificial neural network.