Abstract: A direct speech-to-speech translation (S2ST) model includes an encoder configured to receive an input speech representation that to an utterance spoken by a source speaker in a first language and encode the input speech representation into a hidden feature representation. The S2ST model also includes an attention module configured to generate a context vector that attends to the hidden representation encoded by the encoder. The S2ST model also includes a decoder configured to receive the context vector generated by the attention module and predict a phoneme representation that corresponds to a translation of the utterance in a second different language. The S2ST model also includes a synthesizer configured to receive the context vector and the phoneme representation and generate a translated synthesized speech representation that corresponds to a translation of the utterance spoken in the different second language.

Abstract: Example aspects of the present disclosure are directed to systems and methods that enable weakly-supervised learning of instance segmentation by applying a cut-and-paste technique to training of a generator model included in a generative adversarial network. In particular, the present disclosure provides a weakly-supervised approach to object instance segmentation. In some implementations, starting with known or predicted object bounding boxes, a generator model can learn to generate object masks by playing a game of cut-and-paste in an adversarial learning setup.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for audio-visual speech separation. A method includes: receiving, by a user device, a first indication of one or more first speakers visible in a current view recorded by a camera of the user device, in response, generating a respective isolated speech signal for each of the one or more first speakers that isolates speech of the first speaker in the current view and sending the isolated speech signals for each of the one or more first speakers to a listening device operatively coupled to the user device, receiving, by the user device, a second indication of one or more second speakers visible in the current view recorded by the camera of the user device, and in response generating and sending a respective isolated speech signal for each of the one or more second speakers to the listening device.

Abstract: A direct speech-to-speech translation (S2ST) model includes an encoder configured to receive an input speech representation that to an utterance spoken by a source speaker in a first language and encode the input speech representation into a hidden feature representation. The S2ST model also includes an attention module configured to generate a context vector that attends to the hidden representation encoded by the encoder. The S2ST model also includes a decoder configured to receive the context vector generated by the attention module and predict a phoneme representation that corresponds to a translation of the utterance in a second different language. The S2ST model also includes a synthesizer configured to receive the context vector and the phoneme representation and generate a translated synthesized speech representation that corresponds to a translation of the utterance spoken in the different second language.

Abstract: A computer system for identifying a posture adopted by a subject is disclosed. The computer system includes a computer processor that is configured to execute the following functions: obtain a set of reference pressure images, each of the reference pressure images associated with a known posture; obtain a recorded pressure image of the subject; compare the recorded pressure image of the subject with a candidate pressure image from the set of reference pressure images; and repeat the comparing function until the recorded pressure image with the candidate pressure image.

Abstract: Example aspects of the present disclosure are directed to systems and methods that enable weakly-supervised learning of instance segmentation by applying a cut-and-paste technique to training of a generator model included in a generative adversarial network. In particular, the present disclosure provides a weakly-supervised approach to object instance segmentation. In some implementations, starting with known or predicted object bounding boxes, a generator model can learn to generate object masks by playing a game of cut-and-paste in an adversarial learning setup.

Abstract: A method for image reconstruction includes defining a dictionary including a set of atoms selected such that patches of natural images can be represented as linear combinations of the atoms. A binary input image, including a single bit of input image data per input pixel, is captured using an image sensor. A maximum-likelihood (ML) estimator is applied, subject to a sparse synthesis prior derived from the dictionary, to the input image data so as to reconstruct an output image comprising multiple bits per output pixel of output image data.

Abstract: A method for image reconstruction includes defining a dictionary including a set of atoms selected such that patches of natural images can be represented as linear combinations of the atoms. A binary input image, including a single bit of input image data per input pixel, is captured using an image sensor. A maximum-likelihood (ML) estimator is applied, subject to a sparse synthesis prior derived from the dictionary, to the input image data so as to reconstruct an output image comprising multiple bits per output pixel of output image data.

Abstract: A computer system for identifying a posture adopted by a subject is disclosed. The computer system includes a computer processor that is configured to execute the following functions: obtain a set of reference pressure images, each of the reference pressure images associated with a known posture; obtain a recorded pressure image of the subject; compare the recorded pressure image of the subject with a candidate pressure image from the set of reference pressure images; and repeat the comparing function until the recorded pressure image with the candidate pressure image.

Abstract: A system supporting subject risk analysis and risk event management for detecting possible risk indications and the risk of a subject developing pressure injuries. The system includes monitoring risk events using a pressure sensing apparatus and by recording pressure values at a plurality of pixels of a sensing mat to determine subject's pressure distribution and associated pressure image at any given time. The mapping of pressure sensing elements coordinates of the pressure sensing apparatus to a subject-based coordinate system using applicable transformation functions enables risk analysis and display of subject's pressure distribution maps representing gathered data at different times. Pressure images of the subject's pressure distributions helps in identifying postures adopted by a subject, determining risk of a subject developing pressure injuries and registering possible bed-exit and bed-fall risk events.

Abstract: A pressure wound prevention system comprising at least one management device which comprises at least one user input apparatus, at least one output mechanism, at least one data input for connecting with a pressure sensor assembly and a processor configured to measure elapsed time, to receive pressure data from the data input and to present an output via the output mechanism. The system may be modular. A method for preventing the development of pressure wounds of a subject is also presented.