Abstract: An image processing method, an image processing device, a training method of a neural network, an image processing method based on a combined neural network model, a constructing method of a combined neural network model, a neural network processor, and a storage medium are provided. The image processing method includes: obtaining, based on an input image, initial feature images of N stages with resolutions from high to low, N is a positive integer and N>2; performing, based on initial feature images of second to N-th stages, cyclic scaling processing on an initial feature image of a first stage, to obtain an intermediate feature image; and performing merging processing on the intermediate feature image to obtain an output image. The cyclic scaling processing includes hierarchically-nested scaling processing of N?1 stages, and scaling processing of each stage includes down-sampling processing, concatenating processing, up-sampling processing, and residual link addition processing.

Abstract: The present invention provides an optical imaging lens. The optical imaging lens comprises eight lens elements positioned in an order from an object side to an image side. Through controlling convex or concave shape of surfaces of the lens elements and satisfying at least one inequality, the optical imaging lens may shorten system length with a good imaging quality.

Abstract: An image processing method, an image processing device, a training method of a neural network, an image processing method based on a combined neural network model, a constructing method of a combined neural network model, a neural network processor, and a storage medium are provided. The image processing method includes: obtaining, based on an input image, initial feature images of N stages with resolutions from high to low, where N is a positive integer and N>2, performing, based on initial feature images of second to N-th stages, cyclic scaling processing on an initial feature image of a first stage, to obtain an intermediate feature image; and preforming merging processing on the intermediate feature image to obtain an output image. The cyclic scaling processing includes hierarchically-nested scaling processing of N?1 stages, and scaling processing of each stage includes down-sampling processing, concatenating processing, up-sampling processing, and residual link addition processing.

Abstract: Garbage collection processing in a distributed shared log system includes a client identifying obsoleted log entries for a shared data object. The client sends information associated with the identified obsoleted log entries to a shared log server. The shared log server receives information associated with obsoleted log entries from all clients in the distributed shared log system and uses the information to delete the obsoleted log entries. The shared log server can update a snapshot mark to indicate the earliest time that a snapshot of the shared log can be taken. The snapshot mark can be updated based on the information associated with obsoleted log entries.

Abstract: Embodiments of the disclosure provide a receiver in an optical sensing system. The exemplary receiver includes a movable detector configured to receive optical signals reflected or scattered from an object scanned by the optical sensing system. The receiver further includes an actuator configured to move the movable detector. The receiver also includes a controller configured to determine a plurality of target positions of the movable detector for receiving the optical signals. The controller is further configured to control the actuator to move the movable detector to the plurality target positions according to a movement pattern.

Abstract: Embodiments of the disclosure provide a system for controlling an emission of laser beams using a plurality of scanning patterns by an optical sensing device. The plurality of scanning patterns interleavingly cover a field of view of the optical sensing device. The system includes a controller that is configured to detect an object within a functional distance range from the optical sensing device based on a reflected first laser beam received by the optical sensing device. The first laser beam is emitted towards a first scanning point in a first scanning pattern. The controller is also configured to determine an aperture extending from the first scanning point, and control the optical sensing device to emit a second laser beam towards a second scanning point in a second scanning pattern and skip the scanning points between the first scanning point and the second scanning point in the aperture.

Abstract: Embodiments of the disclosure provide a system for controlling an emission of laser beams by an optical sensing device. The system includes a controller configured to detect an object within a field of view based on a laser beam reflected by the object and received by the optical sensing device. The laser beam is emitted at a current scanning point according to a first laser emission scheme. The controller is also configured to determine a distance of the object from the optical sensing device. The controller is further configured to, in response to the distance being within a functional distance range, control the optical sensing device to emit laser beams according to a second laser emission scheme towards an aperture extending from the current scanning point. The second laser emission scheme reduces a number of scanning points in the aperture compared to the first laser emission scheme.

Abstract: Embodiments of the disclosure provide receivers for light detection and ranging (LiDAR). In an example, a receiver includes a beam reflecting unit comprising a plurality of digital micromirror devices (DMDs). The beam reflecting unit is configured to receive an input laser beam returned from an object being scanned by the LiDAR and reflect the input laser beam by at least one DMD selectively switched to an “ON” state at an operation angle to form an output laser beam towards a detector. The detector is configured to receive the output laser beam.

Abstract: In some examples, an apparatus is provided. The apparatus comprises: an illuminator having an adjustable field of view (FOV), the FOV being adjusted based on setting a direction of propagation of light to illuminate the FOV; a light detector; and a controller configured to: control the illuminator to project the light along a first direction of propagation to illuminate a first FOV; control the illuminator to project the light along a second direction of propagation to illuminate a second FOV; detect, using the light detector, reflected light received from the first FOV and the second FOV to generate one or more detection outputs for a combined FOV including the first FOV and the second FOV; and perform at least one of a detection operation or a ranging operation of an object in the combined FOV based on the one or more detection outputs.

Abstract: Embodiments of the disclosure provide receivers for light detection and ranging (LiDAR). In an example, a receiver includes an acousto-optical (AO) beam deflecting unit configured to receive an input laser beam and a controller configured to cause an acoustic signal to be applied to the AO beam deflecting unit to deflect the input laser beam for a deflection angle and form an output laser beam towards a beam sensor. The deflection angle between the input and the output laser beams is nonzero.

Abstract: Garbage collection processing in a distributed shared log system includes a client identifying obsoleted log entries for a shared data object. The client sends information associated with the identified obsoleted log entries to a shared log server. The shared log server receives information associated with obsoleted log entries from all clients in the distributed shared log system and uses the information to delete the obsoleted log entries. The shared log server can update a snapshot mark to indicate the earliest time that a snapshot of the shared log can be taken. The snapshot mark can be updated based on the information associated with obsoleted log entries.

Abstract: Embodiments of the disclosure provide receivers for light detection and ranging (LiDAR). In an example, a receiver includes a beam converging device and an EO beam deflecting unit. The beam converging device is configured to receive a laser beam from an object being scanned by the LiDAR and form an input laser beam. The EO beam deflecting unit is configured to generate a non-uniform medium having at least one of a refractive index gradient or a diffraction grating, receive the input laser beam such that the input laser beam impinges upon the non-uniform medium, and form an output laser beam towards a photosensor. An angle between the input and the output laser beams is nonzero.

Abstract: Techniques performed by a data processing system for reconstructing a three-dimensional (3D) model of the face of a human subject herein include obtaining source data comprising a two-dimensional (2D) image, three-dimensional (3D) image, or depth information representing a face of a human subject. Reconstructing the 3D model of the face also includes generating a 3D model of the face of the human subject based on the source data by analyzing the source data to produce a coarse 3D model of the face of the human subject, and refining the coarse 3D model through free form deformation to produce a fitted 3D model. The coarse 3D model may be a 3D Morphable Model (3DMM), and the coarse 3D model may be refined through free-form deformation in which the deformation of the mesh is limited by applying an as-rigid-as-possible (ARAP) deformation constraint.

Abstract: Embodiments of the disclosure provide receivers for light detection and ranging (LiDAR). In an example, a receiver includes a beam reflecting unit having a plurality of digital micromirror devices (DMDs). The beam reflecting unit is configured to receive an input light beam returned from an object being scanned by the LiDAR, reflect a signal in the input light beam by at least one first DMD selectively switched to an “ON” state at an operation angle to form an output signal towards a detector, reflect a noise signal in the input light beam away from the detector by at least one second DMD selectively switched to an “OFF” state at a non-operation angle. The receiver also includes the detector configured to receive the output signal.

Abstract: Embodiments of the disclosure provide receivers for light detection and ranging (LiDAR). In an example, a receiver includes a beam converging device, an AO beam deflecting unit, and a beam sensor. The beam converging device is configured to receive a laser beam from an object being scanned by the LiDAR and form an input laser beam. The AO beam deflecting unit is configured to generate a diffraction grating along a propagating direction of an acoustic wave, receive the input laser beam such that the input laser beam impinges upon the diffraction grating, and form an output laser beam towards the beam sensor. An angle between the input and the output laser beams is nonzero.

Abstract: Embodiments of the disclosure provide a receiver in an optical sensing system for receiving a light beam. The receiver includes a first polarizer configured to pass the light beam of a first polarization. The receiver further includes an electro-optical layer coated with patterned transparent electrodes. An electric field is applied to a selected area of the electro-optical layer through the patterned transparent electrodes, and the electro-optical layer changes a portion of the light beam from the first polarization to a second polarization. The receiver also includes a second polarizer configured to selectively pass the portion of the light beam of the second polarization. The receiver additionally includes a detector configured to receive the portion of the light beam output from the second polarizer.

Abstract: Techniques performed by a data processing system for reconstructing a three-dimensional (3D) model of the face of a human subject herein include obtaining source data comprising a two-dimensional (2D) image, three-dimensional (3D) image, or depth information representing a face of a human subject. Reconstructing the 3D model of the face also includes generating a 3D model of the face of the human subject based on the source data by analyzing the source data to produce a coarse 3D model of the face of the human subject, and refining the coarse 3D model through free form deformation to produce a fitted 3D model. The coarse 3D model may be a 3D Morphable Model (3DMM), and the coarse 3D model may be refined through free-form deformation in which the deformation of the mesh is limited by applying an as-rigid-as-possible (ARAP) deformation constraint.

Abstract: Apparatus and methods for determining the distance to an object by detecting a reflected light beam. A light emitter emits a light beam. An optical fiber is mounted to direct the light beam as an output light beam of the optical fiber. An actuator is coupled proximate a distal end of the optical fiber, for moving the distal end of the optical fiber in a desired pattern. Collimation optics are mounted to intercept the output light beam of the optical fiber and collimate the output light beam of the optical fiber. A photodetector is mounted to receive a reflected light beam after reflecting off an object in an external environment. A control system determines a distance to the object based on an elapsed time between emission of the light beam and a detection of the reflected light beam by the photodetector.

Abstract: Embodiments of the disclosure provide a system for determining a laser emitting scheme of an optical sensing device. The system includes a controller communicatively coupled to a laser emitter of the optical sensing device. The controller is configured to perform operations that includes determining a safety distance range in a field of view of the optical sensing device based on a predetermined tolerance value, causing the laser emitter to emit a first laser pulse having a first power-associated value lower than the predetermined tolerance value in the safety distance range, and detecting whether an object is detected in the safety distance range based on the first laser pulse. In response to no object being detected in the safety distance range, the controller causes the laser emitter to emit a second laser pulse having a second power-associated value higher than the first power-associated value.

Abstract: Embodiments of the disclosure provide a light detection and ranging (LiDAR) system. In an example, the LiDAR system includes a laser source, a scanner, a beam deflecting unit, and a controller. The laser source is configured to emit a laser beam towards an object. The scanner is configured to receive a returned laser beam from the object, and deflect the returned laser beam towards a beam deflecting unit to form a first laser beam traveling along a first direction. The first direction deviates from a reference direction by a deviation angle. The beam deflecting unit is configured to receive the first laser beam, and deflect the first laser beam to form a second laser beam towards a photosensor. The controller is configured to dynamically control a deflection angle of the beam deflecting unit to cause the second laser beam to be deflected towards the photosensor to compensate the deviation angle.