Abstract: A knowledge distillation method for compressing an image segmentation model that is performed in a computing device including one or more processors and a memory storing one or more programs executed by the one or more processors includes training a first image segmentation model including a first image encoder, a first image embedding layer, and a first mask decoder, constructing a second image segmentation model including a second image encoder, a second image embedding layer, and a second mask decoder according to preset constraints, and performing knowledge distillation for the second image embedding layer of the second image segmentation model based on the trained first image segmentation model.

Abstract: A method for film coating by pulsed laser deposition with a plasma grating includes: in step 1, providing a substrate and a target material; in step 2, generating a femtosecond pulsed laser beam which is split by a beam splitting module so as to form a plurality of femtosecond pulsed laser sub-beams; in step 3, performing a first excitation on the target material by one of the split femtosecond pulsed laser sub-beams as a pre-pulse after focus, to generate a first plasma; in step 4, synchronizing the rest of the split femtosecond pulsed laser sub-beams as post-pulses to form, after focus, filaments arriving at a surface of the target material simultaneously, to generate the plasma grating; and in step 5, performing a secondary excitation on the target material by the generated plasma grating to generate a second plasma depositing on the substrate to form the film.

Abstract: Disclosed is a method for evaluating the underground hydrogen storage (UHS) capacity in porous media of depleted gas reservoirs using CO2 as cushion gas, including the following steps: S1: obtaining physical parameters required for UHS capacity evaluation based on geological and production data of the target depleted gas reservoir; S2: calculating the cap-rock breakthrough pressure and fault-slip pressure of the UHS based on the cap and fault properties of the target depleted gas reservoir to further estimate the highest operating pressure of the UHS; S3: developing the H2—CO2—CH4 multi-component material balance equation in underground porous media by considering the dissolution of three components in formation water; S4: substituting the parameters obtained from S1 and S2 into the multi-component material balance equation to calculate the proportion of different gas components in the pore space to finally determine the UHS capacity of the target depleted gas reservoir.

Abstract: Disclosed is a method for evaluating the underground hydrogen storage (UHS) capacity in porous media of depleted gas reservoirs using CO2 as cushion gas, including the following steps: S1: obtaining physical parameters required for UHS capacity evaluation based on geological and production data of the target depleted gas reservoir; S2: calculating the cap-rock breakthrough pressure and fault-slip pressure of the UHS based on the cap and fault properties of the target depleted gas reservoir to further estimate the highest operating pressure of the UHS; S3: developing the H2—CO2—CH4 multi-component material balance equation in underground porous media by considering the dissolution of three components in formation water; S4: substituting the parameters obtained from S1 and S2 into the multi-component material balance equation to calculate the proportion of different gas components in the pore space to finally determine the UHS capacity of the target depleted gas reservoir.

Abstract: The present application provides a method and a unified framework system for full-stack autonomous driving planning. The method comprises: acquiring an image of a scene and converting the image into an image feature, and converting the image feature into a bird's-eye view feature map; detecting agents from the bird's-eye view feature map through track query vectors, and continuously tracking the agents; segmenting different types of map elements from the bird's-eye view feature map through map query vectors, and continuously updating the map elements; predicting a future trajectory of each agent using an interaction between the agents and the different map elements; predicting, according to the predicted future trajectory of each agent, an occupancy grid map over multi-steps into the future; and decoding an ego-vehicle query vector to generate a planned path of an ego-vehicle, and optimizing the planned path using the predicted future multi-step occupancy grid map.

Abstract: The present disclosure provides a method of an ultrafast-pulsed laser deposition and a device thereof, wherein a single emitted femtosecond pulse is split, and the split pulses are synchronized in the time domain, and then coupled with each other to form a plasma grating or lattice to excite the target material once; then multiple pulsed lasers are sequentially coupled multiple times with the plasma gratings or lattices to excite the target material multiple times, and the excited target material is deposited and reacted on the substrate to form a thin film.

Abstract: A pulsed laser deposition method is provided. The method includes emitting a plurality of groups of femtosecond pulses, focusing the plurality of groups of femtosecond pulses into a plurality of groups of femtosecond filaments by lenses, and cross-coupling the plurality of groups of femtosecond filaments to form n beams of plasma gratings; exciting a target material by using a first plasma grating; and adjusting angles of the lenses and time delay between a plurality of beams of femtosecond pulses; coupling and splicing a second plasma grating with the first plasma grating along a grating pattern of the first plasma grating, until a nth plasma grating is coupled and spliced with a (n?1)th plasma grating along a grating pattern of the (n?1)th plasma grating to form a plasma grating channel; and exciting the target material by using the plasma grating channel to complete deposition on a substrate.

Abstract: A processing method and apparatus for ultrafast laser deposition of a multilayer film including a diamond-like carbon film, an anti-reflection film and an anti-fingerprint film includes: generating primary plasma by first excitement on a target material with a femtosecond or picosecond pulsed laser beam as a pre-pulse; and generating secondary plasma by second excitement on the target material under plasma grating, formed by allowing two femtosecond pulsed laser beams to intersect at a small include angle for interaction in the primary plasma, for deposition to coat a film.

Abstract: A method for film coating by pulsed laser deposition with a plasma grating includes: in step 1, providing a substrate and a target material; in step 2, generating a femtosecond pulsed laser beam which is split by a beam splitting module so as to form a plurality of femtosecond pulsed laser sub-beams; in step 3, performing a first excitation on the target material by one of the split femtosecond pulsed laser sub-beams as a pre-pulse after focus, to generate a first plasma; in step 4, synchronizing the rest of the split femtosecond pulsed laser sub-beams as post-pulses to form, after focus, filaments arriving at a surface of the target material simultaneously, to generate the plasma grating; and in step 5, performing a secondary excitation on the target material by the generated plasma grating to generate a second plasma depositing on the substrate to form the film.

Abstract: The present invention belongs to the technical field of computer vision, and provides a video semantic segmentation method based on active learning, comprising an image semantic segmentation module, a data selection module based on the active learning and a label propagation module. The image semantic segmentation module is responsible for segmenting image results and extracting high-level features required by the data selection module; the data selection module selects a data subset with rich information at an image level, and selects pixel blocks to be labeled at a pixel level; and the label propagation module realizes migration from image to video tasks and completes the segmentation result of a video quickly to obtain weakly-supervised data. The present invention can rapidly generate weakly-supervised data sets, reduce the cost of manufacture of the data and optimize the performance of a semantic segmentation network.

Abstract: Illustrative embodiments provide a method and system for communicating air traffic control information. An audio signal comprising voice activity is received. Air traffic control information in the voice activity is identified using an artificial intelligence algorithm. A text transcript of the air traffic control information is generated and displayed on a confirmation display. Voice activity in the audio signal may be detected by identifying portions of the audio signal that comprise speech based on a comparison between the power spectrum of the audio signal and the power spectrum of noise and forming speech segments comprising the portions of the audio signal that comprise speech.

Abstract: The embodiments of the present application discloses an image processing method, an image processing apparatus, an image processing device and a computer-readable storage medium.

Abstract: A system is described. The system includes a first region and a second region. The first region is configured to retain particles and has a first characteristic dimension. The second region is communicatively coupled with the first region such that the particles can travel between the first region and the second region. The second region has a second characteristic dimension of not larger than twice a characteristic length for the particles. The second region has a barrier that offers resistance to or interruption in particle movement.

Abstract: The disclosure features systems, devices, and methods of sectioned compaction of composite materials for large parts. An example system includes a mold assembly, including a bottom wall, one or more side walls coupled to the bottom wall, one or more pistons, wherein the bottom wall, the one or more side walls, and the one or more pistons at least partially define a mold cavity, and wherein each of the one or more pistons are independently movable to reduce the volume of the mold cavity; and a plurality of fasteners configured to selectively immobilize the one or more pistons; and a compression actuator configured to sequentially advance the pistons from an initial position to an actuated position to compress a composite material within the mold cavity.

Abstract: An action recognition method and apparatus related to artificial intelligence and include extracting a spatial feature of a to-be-processed picture, determining a virtual optical flow feature of the to-be-processed picture based on the spatial feature and X spatial features and X optical flow features in a preset feature library, where the X spatial features and the X optical flow features include a one-to-one correspondence, determining a first type of confidence of the to-be-processed picture in different action categories based on similarities between the virtual optical flow feature and Y optical flow features, where each of the Y optical flow features in the preset feature library corresponds to one action category, X and Y are both integers greater than 1, and determining an action category of the to-be-processed picture based on the first type of confidence.

Abstract: A method of manufacturing a battery includes introducing a first material to the battery, providing an anode, a cathode and a separator of the battery; and assembling the anode, the separator and the cathode. The first material is configured and arranged to increase the internal impedance of the battery upon mechanical or thermal loading.

Abstract: A micro-displacement amplifying apparatus comprises two sets of asymmetrical amplifying structures; each set of asymmetrical amplifying structure comprises a plurality of asymmetrical amplifying units connected in series by flexible hinges; the asymmetrical amplifying unit is used for amplifying a micro-displacement; the two sets of asymmetrical amplifying structures are in opposite positions and overlap with each other; the input end and output end are coupled to the asymmetrical amplifying unit by a flexible hinge, respectively; the input end is used for inputting the micro-displacement to the asymmetrical amplifying unit, and the output end is used for outputting the amplified displacement; the two contacting input ends are fixed and coupled to each other, and the two contacting output ends are fixed and coupled to each other. The present disclosure further discloses an amplification method of the micro-displacement amplifying apparatus.

Abstract: An action recognition method and apparatus related to artificial intelligence and include extracting a spatial feature of a to-be-processed picture, determining a virtual optical flow feature of the to-be-processed picture based on the spatial feature and X spatial features and X optical flow features in a preset feature library, where the X spatial features and the X optical flow features include a one-to-one correspondence, determining a first type of confidence of the to-be-processed picture in different action categories based on similarities between the virtual optical flow feature and Y optical flow features, where each of the Y optical flow features in the preset feature library corresponds to one action category, X and Y are both integers greater than 1, and determining an action category of the to-be-processed picture based on the first type of confidence.

Abstract: Action recognition methods are disclosed.