公开(公告)号：US20220164652A1

公开(公告)日：2022-05-26

申请号：US17431012

申请日：2020-01-29

Applicant: Nokia Technologies Oy

Inventor： Caglar AYTEKIN ,  Francesco CRICRI

IPC: G06N3/08

Abstract: There is provided an apparatus comprising means for training a neural network, wherein the training comprises applying a loss function configured to increase sparsity of a weight tensor of the neural network and to cause a plurality of non-zero elements of the weight tensor to be substantially equal to each other; and means for entropy coding the weight tensor to obtain a compressed neural network.

公开(公告)号：US20210168395A1

公开(公告)日：2021-06-03

申请号：US16973620

申请日：2019-07-08

Applicant: Nokia Technologies Oy

Inventor： Francesco CRICRI ,  Antti HALLAPURO ,  Miska HANNUKSELA ,  Jani LAINEMA ,  Emre AKSU ,  Caglar AYTEKIN ,  Ramin GHAZNAVI YOUVALARI

IPC: H04N19/50 ,  H04N19/172 ,  H04N19/154 ,  H04N19/14 ,  H04N19/196 ,  H04N19/105 ,  G06N3/04 ,  G06N3/08

Abstract: An apparatus, a method and a computer program product are described comprising: obtaining or receiving video data; providing a current frame and/or one or more previous frames of the obtained or received video data to an input of a neural network; generating a predicted output at an output of the neural network, wherein the predicted output comprises at least one of one or more predicted future frames of the video data and predicted properties of one or more future frames of the video data; determining one or more processing decisions based, at least in part, on the predicted output; and processing the current frame of the video data at least partially according to the one or more processing decisions.

公开(公告)号：US20220164995A1

公开(公告)日：2022-05-26

申请号：US17430987

申请日：2020-01-29

Applicant: Nokia Technologies Oy

Inventor： Caglar AYTEKIN ,  Francesco CRICRI ,  Mikko HONKALA

IPC: G06T9/00 ,  H04N19/15 ,  H04N19/132 ,  H04N19/196 ,  G06N3/08

Abstract: The embodiments relate to a method comprising compressing input data (I) by means of at least a neural network (E, 310); determining a compression rate for data compression; miming the neural network (E, 310) with the input data (I) to produce an output data (c); removing a number of elements from the output data (c) according to the compression rate to result in a reduced form of the output data (me); and providing the reduced form of the output data (me) and the compression rate to a decoder (D, 320). The embodiments also relate to a method comprising receiving input data (me) for decompression; decompressing the input data (me) by means of at least a neural network (D, 320); determining a decompression rate for decompressing the input data (me); miming the neural network (D, 320) with input data (me) to produce a decompressed output data (ï); padding a number of elements to the compressed input data (me) according to the decompression rate to produce an output data (ï); and providing the output data (ï).

公开(公告)号：US20220141455A1

公开(公告)日：2022-05-05

申请号：US17430893

申请日：2020-01-29

Applicant: Nokia Technologies Oy

Inventor： Francesco CRICRI ,  Caglar AYTEKIN ,  Miska HANNUKSELA ,  Xingyang NI

IPC: H04N19/11 ,  H04N19/85 ,  H04N19/159 ,  H04N19/176 ,  G06N3/08 ,  G06N3/04

Abstract: A method comprising: obtaining a configuration of at least one neural network comprising a plurality of intra-prediction mode agnostic layers and one or more intra-prediction mode specific layers, the one or more intra-prediction mode specific layers corresponding to different intra-prediction modes; obtaining at least one input video frame comprising a plurality of blocks; determining to encode one or more blocks using intra prediction; determining an intra-prediction mode for each of said one or more blocks; grouping blocks having same intra-prediction mode into groups, each group being assigned with a computation path among the plurality of intra-prediction mode agnostic and the one or more intra-prediction mode specific layers; training the plurality of intra-prediction mode agnostic and/or the one or more intra-prediction mode specific layers of the neural networks based on a training loss between an output of the neural networks relating to a group of blocks and ground-truth blocks, wherein the ground-truth blocks are either blocks of the input video frame or reconstructed blocks; and encoding a block using a computation path assigned to an intra-prediction mode for the block.

公开(公告)号：US20200311551A1

公开(公告)日：2020-10-01

申请号：US16828106

申请日：2020-03-24

Applicant: NOKIA TECHNOLOGIES OY

Inventor： Caglar AYTEKIN ,  Francesco CRICRI ,  Yat Hong LAM

IPC: G06N3/08 ,  G06N20/10 ,  G06T9/00 ,  G06T3/40

Abstract: A method, apparatus, and computer program product are provided for training a neural network or providing a pre-trained neural network with the weight-updates being compressible using at least a weight-update compression loss function and/or task loss function. The weight-update compression loss function can comprise a weight-update vector defined as a latest weight vector minus an initial weight vector before training. A pre-trained neural network can be compressed by pruning one or more small-valued weights. The training of the neural network can consider the compressibility of the neural network, for instance, using a compression loss function, such as a task loss and/or a weight-update compression loss. The compressed neural network can be applied within a decoding loop of an encoder side or in a post-processing stage, as well as at a decoder side.