Abstract: The present invention relates to a method for pruning a neural network comprising a plurality of neurons, said method comprising: an initialization phase, wherein input information is fetched comprising at least parameters ({wni,bni}) related to said neural network and a dataset (D) representative of a task that said neural network has to deal with, wherein said parameters ({wni,bni}) comprising a weights vector (wni) and/or a bias (bni) related to at least one neuron of said plurality of neurons; a regularization phase, wherein said neural network is trained according to a training algorithm by using said dataset (D); a thresholding phase, wherein an element (wnij) of said weights vector (wni) is put at zero when its absolute value is below a given threshold (T).

Abstract: A method for learned image compression implemented in an autoencoder includes: a) extracting from an image a latent space by the learnable encoder; b) quantizing the latent space by a quantizer to obtain a quantized latent space; c) entropy coding the quantized latent space by an entropy encoder to obtain a bitstream, wherein an entropy model used to encode the latent space is represented by a probability distribution; d) entropy decoding the bitstream by an entropy decoder to obtain an entropy decoded bitstream; e) feeding the entropy decoded bitstream to the decoder; f) recover a reconstructed image by the decoder; g) training the autoencoder via standard gradient descent of the backpropagated error gradient by finding learnable parameters of the learnable encoder and of the decoder that minimize a rate distortion cost function, wherein the entropy encoder is based on a differentiable formulation of a soft frequency counter.

Abstract: A method includes providing a neural network having a set of weights. The neural network receives an input data structure for generating a corresponding output array according to values of the set of weights. The neural network is trained to obtain a trained neural network. The training includes setting values of the set of weights with a gradient descent algorithm which exploits a cost function including a loss term and a regularization term. The trained neural network is deployed on a device through a communication network, and used by the device. The regularization term is based on a rate of change of elements of the output array caused by variations of the set of weights values.

Abstract: A method for learned image compression implemented in an autoencoder including a learnable encoder and a decoder, the method including: a) extracting from an image a latent space by the learnable encoder; b) quantizing the latent space by a quantizer to obtain a quantized latent space; c) entropy coding the quantized latent space by an entropy encoder to obtain a bitstream, wherein an entropy model used to encode the latent space is represented by a probability distribution; d) entropy decoding the bitstream by an entropy decoder to obtain an entropy decoded bitstream; e) feeding the entropy decoded bitstream to the decoder; f) recover a reconstructed image by the decoder; g) training the autoencoder via standard gradient descent of the backpropagated error gradient by finding learnable parameters of the learnable encoder and of the decoder that minimize a rate distortion cost function, wherein the entropy encoder is based on a differentiable formulation of a soft frequency counter.

Abstract: A method of processing a first image in a first plurality of images, wherein the first image is divided into a plurality of pixel blocks, is proposed, which comprises, for a current block of the first image: selecting, in a set of a plurality of predefined interpolation filters, an interpolation filter based on a prediction of an interpolation filter determined by a supervised learning algorithm to which data related to the current block is input; and using the selected interpolation filter for calculating fractional pixel values in a second image of the plurality of images for a temporal prediction of pixels of the current block based on a reference block correlated to the current block in the second image, wherein the second image is distinct from the first image and was previously encoded according to an image encoding sequence for encoding the images of the plurality of images.

Abstract: Method for compressing a sequence of images comprising a first image and a second image, the method comprising the steps of: generating a first descriptor comprising parameters for displaying a computer-generated graphical element in the first image, the graphical element being of non-photographic origin, and the display parameters not comprising pixel values; processing the second image so as to determine an event which gave rise to a potential variation in the parameters for displaying the graphical element between the first image and the second image; generating a second descriptor comprising an event code indicating the determined event.

Abstract: The present invention relates to a method for pruning a neural network comprising a plurality of neurons, said method comprising: an initialization phase, wherein input information is fetched comprising at least parameters ({wni,bni}) related to said neural network and a dataset (D) representative of a task that said neural network has to deal with, wherein said parameters ({wni,bni}) comprising a weights vector (wni) and/or a bias (bni) related to at least one neuron of said plurality of neurons; a regularization phase, wherein said neural network is trained according to a training algorithm by using said dataset (D); a thresholding phase, wherein an element (wnij) of said weights vector (wni) is put at zero when its absolute value is below a given threshold (T).

Abstract: A method of processing a first image in a first plurality of images, wherein the first image is divided into a plurality of pixel blocks, is proposed, which comprises, for a current block of the first image: selecting, in a set of a plurality of predefined interpolation filters, an interpolation filter based on a prediction of an interpolation filter determined by a supervised learning algorithm to which data related to the current block is input; and using the selected interpolation filter for calculating fractional pixel values in a second image of the plurality of images for a temporal prediction of pixels of the current block based on a reference block correlated to the current block in the second image, wherein the second image is distinct from the first image and was previously encoded according to an image encoding sequence for encoding the images of the plurality of images.

Abstract: A method includes providing a neural network having a set of weights. The neural network receives an input data structure for generating a corresponding output array according to values of the set of weights. The neural network is trained to obtain a trained neural network. The training includes setting values of the set of weights with a gradient descent algorithm which exploits a cost function including a loss term and a regularization term. The trained neural network is deployed on a device through a communication network, and used by the device. The regularization term is based on a rate of change of elements of the output array caused by variations of the set of weights values.