Abstract: An artificial intelligence framework is described that incorporates a number of neural networks and a number of transformers for converting a two-dimensional image into three-dimensional semantic information. According to at least one embodiment, neural networks convert one or more images into a set of image feature maps, depth information associated with the one or more images, and query proposals based on the depth information, a first transformer implements a cross-attention mechanism to process the set of image feature maps in accordance with the query proposals, and a second transformer implements a self-attention mechanism to convert the initial voxel features into refined voxel features, which are up-sampled and processed by a lightweight neural network to generate the three-dimensional semantic information.

Abstract: This disclosure provides a deep learning method, apparatus, device, and medium for task joint optimization based on vascular centerline guidance, including: in a process of extracting a global image feature by an encoder, extracting a local image feature after performing data augmentation on a local image with morphological specificity based on the vascular centerline guidance, and fusing the local image feature with the global image feature to form a fused image feature; and forming a bi-directional feature correction process between decoders of different decoding branches. The vascular centerline guidance is used to provide gated feature optimization and computational efficiency adjustment for an encoding process. Specific morphological data of the vascular centerline guidance is used to form a control mechanism for modulating efficiency of local feature extraction. A scarcity problem of hemodynamic annotations has been alleviated and generalization ability of a model has been enhanced.

Abstract: Provided are an upgrade method and apparatus, a device and a storage medium. The method is performed by a target terminal, and the target terminal is connected to a CAN hardware device and a digital key system. The upgrade method includes: acquiring identification information of the CAN hardware device and a target upgrade mode; determining a target upgrade node according to the target upgrade mode; and upgrading the digital key system according to an upgrade file corresponding to the target upgrade node and the identification information of the CAN hardware device. Therefore, other software does not need to be installed, and the operation is convenient and fast, thereby improving the upgrade efficiency; upgrading does not need to be performed based on a source code, thereby protecting the source code from being exposed.

Abstract: This disclosure provides a task joint optimization method, apparatus, device, and medium for vascular morphology and vascular hemodynamic characteristics, including using a hybrid encoder suitable for at least two task types for image feature extraction during an encoding process; configuring a decoder based on a task type to decode an image feature, and forming at least one decoding branch for a vascular morphology task type and at least one decoding branch for a hemodynamic function task type; and in any pair of decoding branches in two task types, forming a bi-directional feature correction process using an output feature of one decoder to correct an output feature of the other decoder between decoders of the two task types. A scarcity problem of hemodynamic annotations has been alleviated and generalization ability and segmentation accuracy of a model have been enhanced.

Abstract: Techniques for emergent scene decomposition from multi-traverse include receiving a plurality of images from multiple traversals of a scene; generating a plurality of 3D Gaussians from the plurality of images; projecting each of the plurality of 3D Gaussians to generate a plurality of rendered 2D images; extracting a feature map from each of the plurality of images and the plurality of rendered 2D images; generating ephemeral objects masks for the plurality of images from the feature maps and the plurality of rendered 2D images; generating optimized 3D Gaussians from the plurality of images, the plurality of rendered 2D images, and the ephemeral objects masks; and generating a 3D environment from the optimized 3D Gaussians.

Abstract: Scene reconstruction is a computer vision process that creates a model of a scene from a given input, usually including creating a three-dimensional (3D) scene model from one or more input two-dimensional (2D) images of the scene. High-quality scene reconstruction and rendering is useful for various applications, such as autonomous agent applications and scene editing applications. Existing scene reconstruction methods encounter limitations in dynamic scenes where moving objects are not consistent across views from different times, and these methods generally lack the motion cues essential for effective object-environment decomposition and further lead to incomplete environment reconstruction of novel views from persistent occlusion of environmental structures.

Abstract: Methods and apparatuses are described for a dual screen extended reality (XR) headset device, including a first display screen viewable by a wearer of the headset device and a second display screen on an external surface of the headset device and viewable by other persons in proximity to the wearer. The headset device generates a virtual environment for display to the wearer via the first display screen. The headset device captures data associated with execution of the software application, data associated with the virtual environment being displayed to the wearer, and data associated with the wearer. The headset device displays a user interface based upon at least a portion of the captured data on the second display screen, and detects input associated with the user interface displayed on the second display screen. The headset device updates the virtual environment and access to software features based upon the detected input.

Abstract: A multi-process parameter optimization method for low-pressure casting of aluminum alloy wheel hubs is disclosed, relating to the technical field of low-pressure casting of automobile wheel hubs. The method includes: building a three-dimensional model of an aluminum alloy wheel hub; setting initial production process parameters; numerically simulating a low-pressure casting process; analyzing temperature distribution in key points of a mold; adjusting the production process parameters based on temperature distribution; repeating the above steps to further optimize the production process parameters and obtain an optimal combination of process parameters; collecting and analyzing actual production data and building a model; and optimizing the process parameters by using a dynamic multi-objective particle swarm.

Abstract: Provided is a method for evaluating the stability of the cooling effect of a cooling system for low-pressure casting of an aluminum alloy wheel hub, relating to the technical field of low-pressure casting of automobile wheel hubs. The method includes: selecting a mold object; arranging a thermocouple, and acquiring temperature data; changing an initial temperature at a temperature measuring point of the thermocouple, and further acquiring real-time temperature data; extracting characteristic temperature data, and performing linear regression between an initial temperature value and a characteristic temperature value; and performing temperature data fluctuation analysis, and quantitatively measuring the stability of the cooling effect of the cooling system by using a maximum value and a minimum value of deviation of discrete points from a fitting curve as indexes.

Abstract: An artificial intelligence framework is described that incorporates a number of neural networks and a number of transformers for converting a two-dimensional image into three-dimensional semantic information. Neural networks convert one or more images into a set of image feature maps, depth information associated with the one or more images, and query proposals based on the depth information. A first transformer implements a cross-attention mechanism to process the set of image feature maps in accordance with the query proposals. The output of the first transformer is combined with a mask token to generate initial voxel features of the scene. A second transformer implements a self-attention mechanism to convert the initial voxel features into refined voxel features, which are up-sampled and processed by a lightweight neural network to generate the three-dimensional semantic information, which may be used by, e.g., an autonomous vehicle for various advanced driver assistance system (ADAS) functions.

Abstract: The present disclosure provides an intelligent temperature control method for a casting system, including: obtaining feature regions of a casting mold and arranging thermocouples in the feature regions of the mold based on structural characteristics of castings; building a random forest model and performing recursive feature elimination based on temperature measurement results of the thermocouples and casting quality inspection results, to determine a correlation between temperature measurement data of each thermocouple and casting quality, thereby optimizing a quantity of the thermocouples and screening the thermocouples; and analyzing temperature data of screened thermocouple temperature measurement points, cooling process parameters, and corresponding casting quality, constructing a relation among the cooling process parameters, an initial temperature of each thermocouple in the mold, and the casting quality through a gradient boosting decision tree model, and controlling the temperature of the casting sys

Abstract: A full-automatic preparation method of a cast-off cell smear includes the following steps: S1: pretreatment of a cast-off cell: automatically pretreating the cast-off cell to be prepared to remove interfering substances; S2: automatically transferring a pretreated cast-off cell to a cyclic settling module to allow constant-temperature settling, and staining and moisturizing the pretreated cast-off cell; and S3: subjecting a moisturized cast-off cell to automatic liquid sealing. The full-automatic preparation method of the cast-off cell smear realizes the automation of smear preparation, staining, and sealing, and also realizes the unattended high-throughput output and the high-quality smear preparation and sealing.

Abstract: An integrated constant-temperature fully-automatic cell smear preparation device includes a rotary settling disc, where the rotary settling disc includes a settling disc base and a turntable body connected to a settling disc body through a fixed shaft; a glass slide addition mechanism arranged adjacent to a glass slide inlet position of the rotary settling disc; a cleaning and sample addition integrated device, where the cleaning and sample addition integrated device is arranged on a side of the rotary settling disc, the cleaning and sample addition integrated device includes a cleaning assembly and a sample addition assembly, and the cleaning assembly and the sample addition assembly add fluid to one or more different settling cups simultaneously as the turntable body rotates intermittently; and a glass slide discharging mechanism arranged adjacent to a glass slide outlet position of the rotary settling disc.

Abstract: Disclosed are a thermoforming device and method for a flexible printed circuit board. The thermoforming device may include: a first forming mechanism provided with a first forming portion, the first forming portion being configured to carry the flexible printed circuit board to be processed; a second forming mechanism provided with a second forming portion which matches with the first forming portion, at least one of the first forming portion or the second forming portion is provided with a corresponding heating module configured for heat treatment of the flexible printed circuit board; a first driving mechanism configured to drive the first forming mechanism and the second forming mechanism to be displaced relative to each other in a first direction; a second driving mechanism configured to drive the second forming mechanism and the first forming mechanism to be displaced relative to each other in a second direction perpendicular to the first direction; and a controller.

Abstract: An integrated circuit (“IC”) chip having fault protection features. The IC chip in an embodiment may include a lead frame having a first portion adapted to accommodate an IC bare die and a second portion separated from first portion and adapted to be coupled to the IC bare die by an electrically conductive connector. The electrically conductive connector includes a connector bridging portion having a reduced size in comparison with a remainder of the electrically conductive connector and having an elevated portion that is protruding relative to the remainder of the electrically conductive connector.

Abstract: An integrated circuit (“IC”) chip having fault protection features. The IC chip in an embodiment may include a lead frame having a first portion adapted to accommodate an IC bare die and a second portion separated from first portion and adapted to be configured as an electrically conductive lead pad. The electrically conductive connector includes a lead fuse section having a reduced size in comparison with a remainder of the second portion.

Abstract: The present invention discloses a multistage variable burning rate solid rocket motor and a forming method. A multi-material composite fiber skeleton structure is embedded inside a propellant in the present invention, so that the original burning rate of the propellant presents multistage changes to achieve the function of multistage thrust. The fiber skeleton structure embedded in the solid rocket motor is formed by additive manufacturing, fixed in a combustion chamber, and then charged at one time to complete the forming of a motor grain. The design can greatly increase the burning rates of the propellants of end-burning and bore-burning motors, and can also conduct multistage adjustment for the burning rates of the motors by changing the fiber skeleton and structure, so as to satisfy the requirements of missile weapon systems for multistage power propulsion. The present invention greatly simplifies the technological production process of the existing single-chamber multi-thrust motor.

Abstract: Methods and systems for automated testing of virtual reality software applications using robotic devices include a server that converts a test workflow description into a sequence of actions in a format interpretable by a robotic device. The server transmits the sequence of actions to the robotic device. For each action in the sequence of actions, the server a) receives, from the robotic device, execution data corresponding to an outcome of execution of the action by the robotic device; b) captures, from a virtual reality device, one or more of image data and vector data corresponding to a current state of a virtual reality software application; c) analyzes one or more of the image data, the vector data, and the execution data to determine whether the robotic device successfully executed the action; and d) upon determining that the robotic device successfully executed the action, repeats steps a)-d) for the next action.

Abstract: A charging unit adapted to be used for driving an integrated semiconductor laser device. The charging unit may include a first input terminal adapted to receive an input driving energy, a second input terminal adapted to be configured to enable operation or disable operation of the charging unit, a third input terminal adapted to be configured to receive a charging control PWM signal, and a second plurality of output terminals. The charging unit may be adapted to be configured to switch ON or switch OFF the second plurality of output terminals based on the charging control PWM signal PWM_C in a predetermined pattern.