238 results about "Projection plane" patented technology

The invention relates to the technical field of deep neural network training, and discloses a deep neural network training diagnosis method and system based on gradient visualization and loss curved surface analysis, and the method comprises the steps: obtaining a parameter snapshot of a neural network trained to t-th iteration, sampling k random detection vectors, and calculating a Hessian vector product; performing Lanczos iteration on each random detection vector to obtain k symmetric tridiagonal matrixes and calculating a characteristic value density spectrum; mapping the k symmetric tridiagonal matrixes into orthogonal projection base vectors, constructing a two-dimensional projection plane, determining a saddle point state according to the integral of the characteristic value density spectrum in a negative value region, determining a flat region according to the peak value of the characteristic value density spectrum in a zero value region, and determining a sharp minimum value according to the peak value of the characteristic value density spectrum in a positive value region; according to the method, the descending path with the steepest curvature can be captured, and the dominant mode of the dynamics can be optimized, so that the projected loss curved surface truly reflects the geometric structure change in the training process, and the pathological diagnosis of the training process based on the second-order geometric information is realized.

A large-scale point cloud-oriented two-dimensional regularized planar projection method and an encoding and decoding method are disclosed. The two-dimensional regularized planar projection method includes: obtaining original point cloud data; initializing a two-dimensional projection plane structure of a point cloud; and determining a mapping relationship between the original point cloud data and the two-dimensional projection plane structure to obtain a two-dimensional regularized projection plane structure corresponding to the point cloud. The encoding and decoding method includes: predicting based on the two-dimensional regularized projection plane structure of the point cloud to obtain to-be-encoded data; dividing the to-be-encoded data into a first type of to-be-encoded data and a second type of to-be-encoded data; and encoding different data in the first type of to-be-encoded data separately in a preset encoding manner, and correspondingly encoding the second type of to-be-encoded data, to obtain a geometric information bitstream.

A system for presenting digital effects in a head-wearable apparatus is provided. The system generates, by a digital effect application, one or more digital effects for presentation on a digital effect display component and receives, by the digital effect display component, the one or more digital effects, the digital effect display component. The system receives, by the digital effect display component from the digital effect application, positioning information for the one or more digital effects in three-dimensional (3D) space. The system configures, by the digital effect display component, a two-dimensional (2D) re-projection plane based on the positioning information received from the digital effect application and displays, by the digital effect display component, the one or more digital effects on the 2D re-projection plane.

The invention discloses an unmanned aerial vehicle image intelligent perception plotting method and system based on view GIS fusion. The method comprises the steps that a sequence image of a target area is collected through an unmanned aerial vehicle platform; target detection, classification and tracking are carried out based on deep learning, and target pixel coordinates are identified; constructing an intelligent mapping model from a pixel coordinate system to a geographic coordinate system to realize geographic positioning of a target; precise space calibration is carried out on the illegal target; and in combination with space calibration, interactive plotting and visualization of target illegal behaviors are realized at a WebGIS end, and a structured law enforcement evidence packet is formed. According to the method, the POS data and the visual information are deeply fused, the dynamic projection plane estimation technology is adopted, the geographic positioning precision better than that of a traditional method is achieved, the method has obvious advantages especially in complex environments such as urban areas, and high positioning precision can be achieved.

This application discloses a method, apparatus, device, storage medium, and program product for screening face images, belonging to the field of face image screening technology. The method includes: acquiring multiple face images; detecting facial key points and target facial regions from the face images; calculating a first three-dimensional pose angle of the face images based on the facial key points; performing error compensation correction on the first three-dimensional pose angle to obtain a second three-dimensional pose angle; mapping the coordinates of the key points of the target facial region to a virtual projection plane when the face is facing the camera, based on the second three-dimensional pose angle, to obtain corrected coordinates of the key points of the target facial region; calculating state parameters of the target facial region based on the corrected coordinates of the key points of the target facial region; and selecting face images from the multiple face images whose target facial region state parameters satisfy a first preset condition and whose second three-dimensional pose angle satisfies a second preset condition, thereby obtaining a target image. According to the embodiments of this application, the accuracy and efficiency of the screening results can be improved.

Embodiments of the present disclosure relate to the technical field of computers, and provide a cylindrical surface detection method and apparatus. The method comprises: obtaining a plane detection result corresponding to point clouds to be detected; dividing approximately parallel planes in the plane detection result into plane groups; for each plane group, determining, from said point clouds, a point cloud set corresponding to the plane group and a projection plane corresponding to the point cloud set, wherein point cloud points in one point cloud set and an acquisition device configured to obtain said point clouds are located on the same side of the projection plane corresponding to the point cloud set, and the projection plane is a plane in the corresponding plane group; determining a target point cloud set and a target projection plane from the obtained point cloud sets and the corresponding projection planes on the basis of the projection point distribution of the point cloud sets on the corresponding projection planes; and detecting a target cylindrical surface on the basis of the target point cloud set and the target projection plane. In this way, the cylindrical surface can be detected from the point clouds, and high detection accuracy is achieved.

The utility model discloses a lens vibration eccentricity detection device relates to lens eccentricity detection technical field. The device includes: light source generator, is fixed in the horizontal surface of detection platform, and the object side of measured lens is fixedly connected with light source generator, projection plane is set up in the image side of measured lens, the projection plane is equipped with positioning figure, is used for the optical signal of light source generator and forms the projection on the projection plane through measured lens, camera, with projection plane opposite setting, is used for real -time shooting the image on projection plane, including the optical signal image of measured lens still time and the optical signal image of measured lens when being subjected to the vibration, terminal equipment is connected with camera, obtains the image of camera shooting, through comparing the optical signal image of measured lens still time and the optical signal image of measured lens when being subjected to the vibration, detects the eccentricity when the lens vibration.

The refrigeration equipment comprises a box body and a door body, the door body comprises a door body and a decoration plate assembly, the door body comprises vacuum glass, the decoration plate assembly is arranged on the side, away from the box body, of the door body, the decoration plate assembly comprises a door plate and a decoration strip, and the decoration strip is arranged on a projection plane perpendicular to the direction from the door body to the box body. The door plate and the decoration strip surround the vacuum glass, one side of the decoration strip is connected with the door plate through a connecting piece, and the other side of the decoration strip abuts against the vacuum glass. The decoration strip makes up various installation gaps, no matter what kind of errors are generated in the installation process, the decoration strip can fill the gaps through flexible adjustment, and therefore visual consistency of products in different installation environments is guaranteed, the attractiveness problem of a door body is remarkably solved, a more flexible installation mode is provided for a user, and the installation efficiency is improved. And the refrigeration equipment has higher competitiveness in the high-end market.

An Artificial Intelligence Gaze Interface Sensor (AI GIS) system and method for precisely tracking human gaze on a digital or projected view plane without head-mounted hardware. A context camera detects a user's face and matches it against stored biometric profiles using AI facial recognition, automatically loading personalized eye geometry, inter-pupillary distance and gaze calibration data. A conditional activation controller initializes a second Peregrine Foveal Camera (PFC), modeled after the dual-fovea retina of the Falco peregrinus, only upon confirmed eye detection. An AI algorithm triggers visual effects, hidden content, animations, gaze-based engagements and semantically mapped audio on the digital or projected plane based on gaze coordinate outputs, from a dedicated computational pipeline, correlated to the digital or projected plane. This invention serves as the sensing layer for U.S. application Ser. No. 18 / 395,290.

The utility model relates to the technical field of air conditioners, in particular to a side-blowing outdoor unit. The side-blowing outdoor unit comprises an axial flow impeller, a flow guide ring and a flow guide ring; the flow guide ring surrounds the air outlet side of the axial flow impeller, an air inlet and an air outlet of the flow guide ring are both in a horn shape, the flow guide ring is arranged on the air outlet side of the flow guide ring, the cross section of the flow guide ring is in an arc shape, and the concave side of the flow guide ring faces the flow guide ring. The projection of the inner edge of the flow guide ring on the target projection surface is a first projection, the projection of the inner wall of the narrowest part of the flow guide ring on the target projection surface is a second projection, the first projection is located on the periphery of the second projection, and the target projection surface is perpendicular to the axis of the axial flow impeller. According to the side-blowing outdoor unit provided by the utility model, the air outlet side of the flow guide ring is provided with the flow guide ring which meets specific requirements, so that the turbulence of airflow near the air outlet of the flow guide ring can be inhibited, the efficiency of the motor can be improved, and the noise can be reduced.

The invention belongs to the technical field of underwater sound channels and physical fields, and relates to a numerical mixing calculation method for a forward scattering sound field of an underwater regular-shape object. The method comprises the following steps: acquiring an initial projection contour of a three-dimensional geometric model of a regular shape object in a specified projection plane and an incident direction; discrete points of the initial projection contour are divided into upper boundary points and lower boundary points; performing interpolation completion on the upper boundary points and the lower boundary points, so that the upper boundary points and the lower boundary points are the same in number and share abscissas; discretizing the supplemented projection contour by using a trapezoidal plane element; calculating a forward scattering far field morphological function based on a discretization result; and calculating the forward scattering target intensity of the object. The method provided by the invention can cover the shape range of the scattering object suitable for the analysis method, does not need to calculate special functions such as Bessel and spherical harmonic and a complex series expression formed by combining the special functions, and obviously improves the calculation efficiency.

An information processing apparatus includes a processing unit configured to create first two-dimensional data based on a first three-dimensional point cloud and a projection plane and create second two-dimensional data based on a second three-dimensional point cloud and the projection plane, and a registration unit configured to register the first three-dimensional point cloud and the second three-dimensional point cloud through relative movement of the first two-dimensional data and the second two-dimensional data.

The present technology provides an illumination device including a light source section, a phase modulation section, and a control section. The light source section has a light emitting element. The phase modulation section performs spatial light phase modulation on incident light from the light source section. The control section controls the phase modulation section so that a plurality of domains formed by dividing a phase modulation plane of the phase modulation section reproduce, in a common area on a projection plane, a light intensity distribution based on a common phase distribution determined based on a free-form method.

The invention relates to the technical field of projectors, and discloses a picture calibration method and device, a projector and a medium. The method comprises the following steps: acquiring attitude angle information and positioning and orientation information of a projection lens; the positioning and orientation information represents distances and directions of the projection lens relative to at least three non-collinear reference points in a projection plane; determining three-dimensional included angle information between the projection plane and the optical axis of the projection lens based on the attitude angle information and the positioning and orientation information; performing angle calculation on the LCD screen based on the three-dimensional included angle information so as to calculate a target pitch angle value and a target deflection angle value which the LCD screen needs to reach; and driving the screen adjusting component to adjust the angle of the LCD screen by taking the pitch angle and the deflection angle of the LCD screen respectively reaching the target pitch angle value and the target deflection angle value. According to the embodiment of the invention, the projection picture quality of the single-chip LCD projector during side projection can be improved.

The present application relates to the technical field of projectors, and specifically provides a kind of intelligent projector control method based on color correction and depth correction, image and depth information of projection plane are collected by optical perception device, and color deviation tensor and depth deviation tensor between it and ideal plane are calculated, color correction and depth correction are carried out on pre-projection image according to color deviation tensor and depth deviation tensor, so that projected image will not be affected by color and flatness on projection plane.The present application avoids color interference and flatness influence of projection plane, ensures that projector can realize flat, achromatic projection under visual effect in extreme conditions such as plane and space irregularity, color stain and the like, and solves the problems such as precision limitation, single scene and high cost in existing projection mapping application.

The application relates to the technical field of CAD model processing, in particular to a triangular mesh quality optimization method and system based on local reconstruction, which comprises the following steps: acquiring a triangular mesh, and screening a first triangular mesh to be optimized based on a preset regularity; performing basic topological adjustment on the first triangular mesh to be optimized, and obtaining a second triangular mesh to be optimized; acquiring a first local neighborhood mesh of the second triangular mesh to be optimized, moving the vertexes of the triangular mesh in the first local neighborhood mesh, and obtaining a third triangular mesh to be optimized; acquiring a second local neighborhood mesh of the third triangular mesh to be optimized and a projection plane of the second local neighborhood mesh, determining a layout triangle according to the area of the projection plane and a target side length, and mapping the layout triangle to the second local neighborhood mesh. The application can optimize the quality of the triangular mesh on the basis of retaining the original geometric characteristics of the triangular mesh, and greatly improve the calculation efficiency of the triangular mesh optimization.

This invention belongs to the field of medical signal processing technology, specifically a method and prediction system for constructing a three-dimensional tumor motion trajectory prediction model. It involves collecting temporal data of point clouds from the human chest and abdomen, as well as temporal location information of the tumor, and establishing a unified spatial coordinate system. The temporal data of the point clouds are divided into n time periods of total duration T1, and the temporal location information of the tumor is divided into n time periods of total duration T2, resulting in n training samples. The point clouds of the training samples are further divided into several pseudo-voxel sub-blocks according to the human projection plane, and K pseudo-voxel sub-blocks are selected based on motion saliency to obtain the temporal feature vector of the tumor motion. The temporal feature vector of the tumor three-dimensional motion trajectory prediction is used as input, and the corresponding tumor temporal location information is used as output to train an attention-enhanced dilated convolutional prediction network. This invention enables high-precision, non-invasive prediction of the three-dimensional tumor motion trajectory.

An image capturing apparatus 1 has a barrel unit 2 and a tripod screw portion 35a, and includes a tripod attaching portion 35 disposed with respect to the barrel unit 2 in a Y direction orthogonal to a Z direction as an optical axis direction, a battery 80 disposed between the barrel unit 2 and the tripod attaching portion 35 in the Y direction, and a stand 9 which is held rotatable about a rotational axis A2 parallel to an X direction orthogonal to the Z direction and the Y direction and has a connection portion connecting leg portions provided at ends in the X direction. The battery 80 is plate-like and has a rectangular parallelepiped shape, and at the same time is disposed such that a thickness direction thereof is parallel to the Z direction. The tripod screw portion 35a is exposed to appearance in a bottom surface of the image capturing apparatus 1. The rotational axis A2 overlaps the tripod screw portion 35a on a projection plane as viewed in the X direction, and at the same time, overlaps the tripod screw portion 35a, the battery 80, and the barrel unit 2 on a projection plane as viewed in the Y direction.

The present application relates to the technical field of measuring the precision of gaze estimation algorithm, and proposes a method for measuring the precision of gaze estimation algorithm and a matching device thereof. The device comprises a three-degree-of-freedom motion controller for controlling a point laser emitter to move in the front-back direction opposite to the center of a projection plane, to rotate in the horizontal (yaw) direction, and to rotate in the vertical (pitch) direction. The device further comprises a point laser emitter for irradiating a gaze point on the projection plane, a projection plane for displaying the irradiation point of the point laser emitter, and a camera for capturing an eye image of a subject when the subject gazes at the gaze point. The method can satisfy the precision measurement of the gaze estimation algorithm in the positive and negative directions of pitch and yaw within any eye rotation angle range, can simultaneously satisfy the precision measurement of the gaze estimation algorithm for a larger gaze angle, and can satisfy the precision measurement requirements of the gaze estimation algorithm in different distance ranges.

An image projection system (100; 200), comprising: a first transmitter (212d) configured to generate first light rays corresponding to a first projection plane and to send the first light rays along a first transmission path; a first collimation lens (213d) arranged on the first transmission path, the first collimation lens being configured to receive the first light rays and to produce first collimated light rays to be projected onto an eye to reproduce a first projection image that is perceived on the first projection plane; a second transmitter (212i) configured to generate second light beams corresponding to a second projection plane different from the first projection plane and to transmit the second light beams along a second transmission path; a second collimating lens (213i) arranged on the second transmitting path, the second collimating lens being configured to receive the second light rays and to produce second collimated light rays to be projected onto the eye to reproduce a second projection image that is perceived on the second projection plane; a first beam combiner (214) arranged at an intersection of the first transmission path and the second transmission path, wherein the first beam combiner is configured to transmit the first collimated light beams and the second collimated light beams on a combined transmission path; and a scanner (225) arranged on the combined transmission path, the scanner being configured to receive the first collimated light beams and the second collimated light beams and to direct the first collimated light beams and the second collimated light beams according to a scanning pattern to reproduce the first projection image and the second projection image to the eye.

The invention discloses a mandibular canal segmentation method and system, electronic equipment and a storage medium. The method comprises the following steps: acquiring an oral maxillofacial region image; cutting the oral maxillofacial region image to obtain a plurality of local images of the mandible; inputting the plurality of local images into a mandibular canal segmentation model, and outputting to obtain a segmentation result; the mandibular canal segmentation model comprises an encoder, a decoder and a feature fusion module, three two-dimensional convolution branches are configured on three orthogonal projection planes in the feature fusion module, and each two-dimensional convolution branch can obtain boundary information, center trend and local bending features of the mandibular canal on different orthogonal planes and fuse the boundary information, the center trend and the local bending features to obtain multi-view features; and fusing the multi-view feature and the local texture feature to obtain an enhanced feature, fusing the local texture feature and the enhanced feature by a decoder, and outputting to obtain a segmentation result. And through the feature fusion module, comprehensive coverage of multi-view geometric features is realized, and the problems of segmentation fracture and boundary blur caused by single feature extraction are effectively avoided.

The utility model provides a data line assembly and a portable mobile power supply. The data line assembly comprises a housing, a data line body and a limiting member. The shell is provided with a containing groove and a first opening communicated with the containing groove, and the axial direction of the first opening is the first direction. The data line body comprises mutually connected wires and a first plug, and the first plug is accommodated in the accommodating groove. The limiting piece is movably connected to the shell, the orthographic projection of the limiting piece on a projection plane perpendicular to the first direction is a first projection, the orthographic projection of the first plug on the projection plane is a second projection, and when the limiting piece is located at the first position, the first projection and the second projection are at least partially overlapped; the first projection and the second projection are arranged at intervals, so that the first plug can extend out of the accommodating groove from the first opening. According to the scheme of the utility model, through position switching of the limiting member, the housing can accommodate the first plug more stably, and the first plug can be taken and placed more conveniently.

The application relates to the technical field of projectors, and particularly provides an intelligent projector control method based on color avoidance and depth correction, candidate regions on a pre-projection image are determined according to a color deviation tensor of a projection plane and a global color fitting plane, a projection region is divided in the candidate regions according to the shape of the pre-projection image, depth correction is performed on the projection region according to a depth deviation tensor of the projection region and a global depth fitting plane, and the pre-projection image after the depth correction is projected on the projection region obtained through color avoidance division. The application avoids the influence of color interference and flatness, guarantees that the projector can realize flat and achromatic projection under visual effect in the case of plane and space irregularity, color stains and other extreme conditions, and solves the problems of precision limitation, single scene and high cost in the existing projection mapping application.

The present disclosure provides an image reconstruction method, device, electronic equipment and storage medium, the image reconstruction method comprising: acquiring a compressed image of a detection object with respect to a projection plane; processing at least one of two-dimensional position information, texture information and depth information included in the compressed image based on a conversion relationship to obtain three-dimensional information, the conversion relationship being determined according to extreme position information of the two-dimensional position information with respect to the projection plane or a projection angle of the two-dimensional position information with respect to the projection plane; and generating point cloud data of the detection object according to a plurality of three-dimensional information.

The utility model provides a laser-assisted ventricular puncture device. The laser-assisted ventricular puncture device comprises a guide core part and a laser transmitter, the guide core part comprises a guide core and a handheld part arranged at the tail end of the guide core, the handheld part is provided with two fixing frames, the two fixing frames are perpendicular to each other on a projection plane perpendicular to the guide core, and the intersection point is coaxial with the guide core; the number of the laser emitters is two, and the two laser emitters are fixed to the two fixing frames respectively. Accurate and visual alignment can be realized; deviation caused by visual errors is avoided.

A linear actuator device including a lens assembly that has a lens, a linear actuator that linearly drives the lens assembly in a linear actuation direction, and a linear guide that has a main guide bar for supporting the lens assembly in a state movable in the linear actuation direction. The lens assembly has two main engagement portions which are arranged with a certain distance in the linear actuation direction and engaged with the main guide bar. The linear actuator is arranged between the two main engagement portions and at the same time arranged in a position overlapping the lens on a projected plane viewed from a direction orthogonal to the linear actuation direction, to apply thrust to the lens assembly in a direction substantially coinciding with the linear actuation direction.

The application discloses a projection figure distortion prediction compensation correction method of a laser scanning projection system, and comprises the following steps: carrying out laser scanning projection system projection; acquiring projection figure distortion data, establishing a neural network training data set and a test data set; establishing an initial topological structure, weights and thresholds of a BP neural network; optimizing the weights and thresholds of the BP neural network based on a particle swarm optimization algorithm; carrying out deviation compensation on the coordinate positions of each projection point according to the training output of the PSO-BP neural network; selecting feature points in each part of a to-be-projected plane; projecting the feature points by using a distortion deviation correction projector based on the PSO-BP neural network; acquiring feature point projection coordinate values; subtracting the ideal projection point coordinates without distortion from the corrected feature point projection coordinates; judging whether the difference value meets an error interval; if yes, it is determined that the projection coordinate value distortion error has been corrected and compensated; if no, returning to step three for feedback type multiple iterations until the difference value meets the error interval.

An image acquisition device has an optoelectronic image sensor, a position sensor, a graphical user output interface and an evaluation unit. The image sensor comprises an image data set representing an imaging of a scene located in front of the image sensor on an image plane of the image sensor. The position sensor detects a spatial position of the image plane relative to a reference direction and provides position data indicating a rotation angle and a tilt angle. The evaluation unit determines a projection transformation using the position data, which maps the image data set onto a projection plane tilted relative to the image plane according to the rotation and the tilt. The evaluation unit determines an image portion of the image data set in the projection plane mapped onto the projection plane by the projection transformation and displays the image portion and an area of the imaged scene located at least within the image portion in the graphical user output interface.

The invention discloses an indoor three-dimensional model generation method and device, equipment, a medium and a product, and the method comprises the steps: obtaining indoor three-dimensional scene data which comprises a plurality of objects, and each object carries corresponding identification information; projecting each object in the three-dimensional scene data to a projection plane based on a preset projection direction, and generating a scene top view comprising a plurality of two-dimensional blocks; receiving a modification operation of a user for the scene top view; on the basis of the modification operation, updating depth information of the modified target two-dimensional block, and generating a new scene top view comprising the hierarchical display control on the basis of the updated depth information; and determining the corresponding three-dimensional model based on the identification information of the object corresponding to each two-dimensional block in the new scene top view, and synthesizing the plurality of three-dimensional models into the indoor three-dimensional model, so that the generation efficiency of the indoor three-dimensional model can be improved through the mode.

The application belongs to the technical field of medical signal processing, and particularly relates to a construction method and a prediction system of a tumor three-dimensional motion trajectory prediction model. The body surface point cloud time series data and tumor time series position information of the human chest and abdomen are collected, and a unified space coordinate system is established. The body surface point cloud time series data is divided into n time periods with a total length of T1, and the tumor time series position information is divided into n time periods with a total length of T2, thereby obtaining n training samples. The body surface point cloud of the training sample is divided into a plurality of pseudo-voxel sub-blocks according to the human projection plane, and K pseudo-voxel sub-blocks are screened out according to the motion saliency, thereby obtaining a body surface motion time series feature vector. The body surface motion time series feature vector of the tumor three-dimensional motion trajectory prediction is taken as the input, the corresponding tumor time series position information of the tumor three-dimensional motion trajectory prediction is taken as the output, and the attention-enhanced dilated convolution prediction network is trained. The application can realize high-precision and non-invasive tumor three-dimensional motion trajectory prediction.