Abstract: This disclosure provides methods, devices, and systems for improving the environmental awareness of a user of a wearable audio device. The present implementations more specifically relate to environment-aware signaling for wearable audio devices. In some aspects, a wearable audio device may include one or more speakers configured for audio playback, one or more microphones configured to detect sounds from the surrounding environment while the audio is concurrently being played back, and an environment awareness controller configured to record the sounds detected by the microphones and control one or more outputs of the wearable audio device based, at least in part, on the recorded sounds. More specifically, the environment awareness controller may activate or adjust such outputs to alert a user of the wearable audio device about the recorded sounds responsive to detecting a trigger condition associated with the alert.

Abstract: This disclosure provides methods, devices, and systems for object detection. The present implementations more specifically relate to techniques for improving distant object detection in memory-constrained computer vision systems. In some aspects, a computer vision system may include an ROI extraction component, a feature pyramid network (FPN) having a number (N) of pyramid levels, and N network heads associated with the N pyramid levels, respectively. The FPN extracts N feature maps from an input image, where the N feature maps are associated with the N pyramid levels, respectively, and each of the N network heads performs an object detection operation on a respective feature map of the N feature maps. The ROI extraction component selects a region of the feature map associated with the lowest pyramid level for distant object detection so that the object detection operation performed on that feature map is confined to the selected region.

Abstract: A method for controlling a bias of a neural network, the method may include training the neural network by using a loss function that is responsive to classes of a classification process and sensitivity input values thereby setting the bias.

Abstract: This disclosure provides methods, devices, and systems for machine learning. The present implementations more specifically relate to techniques for reducing the computational load of a convolutional neural network (CNN) when processing successive video frames. In some aspects, a machine learning system may cache or store the outputs (also referred to as “activations”) produced by one or more layers of a CNN so that one or more cached activations can be substituted for respective activations that would otherwise be computed by the CNN when processing a subsequent video frame. For example, the machine learning system may compare each video frame with a preceding frame of the video to detect pixels that undergo significant changes between successive frames (also referred to as “motion pixels”). In some aspects, the CNN may only perform neural network operations that involve one or more motion pixels or features derived from a motion pixel.

Abstract: A first mobile communication device that includes a first microphone, a first speaker, and a first delay unit. The first microphone is configured to (i) receive, during a conference call, a first user first microphone signal from a first user, and (ii) output a first microphone digital signal to the first delay unit. The first user first microphone signal represents audio content outputted by the first user. The first delay unit is configured to delay, by a delay period, the first microphone digital signal to provide a delayed first user first device digital signal. The first mobile communication device is configured to output, to a mixer, the delayed first user first device digital signal. The delay period is determined based on measurements executed by at least one mobile communication device out of the first mobile communication device, a second mobile communication device and a third mobile communication device.

Abstract: A first mobile communication device that includes a first microphone, a first speaker, and a first delay unit. The first microphone is configured to (i) receive, during a conference call, a first user first microphone signal from a first user, and (ii) output a first microphone digital signal to the first delay unit. The first user first microphone signal represents audio content outputted by the first user. The first delay unit is configured to delay, by a delay period, the first microphone digital signal to provide a delayed first user first device digital signal. The first mobile communication device is configured to output, to a mixer, the delayed first user first device digital signal. The delay period is determined based on measurements executed by at least one mobile communication device out of the first mobile communication device, a second mobile communication device and a third mobile communication device.

Abstract: A method for distributed neural network processing, the method may include detecting, by a local neural network that belongs to a local device, and based on sensed information, an occurrence of a triggering event for executing or completing a classification or detection process; sending to a remote device, a request for executing or completing the classification or detection process by a remote device that comprises a remote neural network; wherein the remote neural network has more computational resources than the local neural network; determining by the remote device whether to accept the request; and executing or completing, by the remote device, the classification or detection process when determining to accept the request; wherein the executing or completing involves utilizing the remote neural network.

Abstract: There may be provided a method for proximity sensing by a computerized device, the method may include transmitting, by at least one transmitter of a computerized device and during at least one transmission window, one or more transmitted signals, the one or more transmitted signals comprise a transmitted ultrasonic signal; operating at least one receiver of the computerized device to receive, during at least one reception window, one or more received signals that were reflected or scattered due to the transmitting; processing the received signals to provide a processing result, when receiving the received signals by the receiver during the at least one reception window; and determining a proximity of one or more objects to the computerized device based on at least one out of (a) an absence of received signals during the at least one receive window, and (b) the processing results

Abstract: A method for distributed neural network processing, the method may include detecting, by a local neural network that belongs to a local device, and based on sensed information, an occurrence of a triggering event for executing or completing a classification or detection process; sending to a remote device, a request for executing or completing the classification or detection process by a remote device that comprises a remote neural network; wherein the remote neural network has more computational resources than the local neural network; determining by the remote device whether to accept the request; and executing or completing, by the remote device, the classification or detection process when determining to accept the request; wherein the executing or completing involves utilizing the remote neural network.

Abstract: Provided is a wearable audio input/output device. The wearable audio input/output device is provided with a frame, a first audio collection unit, a second audio collection unit, and a first speaker. The frame can be mounted to a user. The first audio collection unit is provided on the frame. The second audio collection unit is provided on the frame. The second audio collection unit is arranged such that the distance between the first audio collection unit and mouth of the user is less than the distance between the second audio collection unit and the mouth of the user. The first speaker is provided on the frame between the first audio collection unit and the second audio collection unit.

Abstract: Provided is a wearable audio input/output device. The wearable audio input/output device is provided with a frame, a first audio collection unit, a second audio collection unit, and a first speaker. The frame can be mounted to a user. The first audio collection unit is provided on the frame. The second audio collection unit is provided on the frame. The second audio collection unit is arranged such that the distance between the first audio collection unit and mouth of the user is less than the distance between the second audio collection unit and the mouth of the user. The first speaker is provided on the frame between the first audio collection unit and the second audio collection unit.