Method and device for partitioned synchronous rendering of a large painting for appreciation

Abstract

The application discloses a large-size painting appreciation picture partition synchronous rendering method and device, and relates to the technical field of human-computer interaction.The method comprises the following steps: using an infrared motion capture system to collect six-degree-of-freedom motion data, generating a three-dimensional line-of-sight feature set through head posture calculation and line-of-sight mapping; according to the three-dimensional line-of-sight feature set, calling an improved PSO algorithm, and performing iterative optimization through chaotic sequence initialization, convergence factor improvement and escape repulsion evolution to obtain an optimal screen segmentation and computing power weight scheme; based on the optimal screen segmentation and computing power weight scheme, performing timing smoothing and heterogeneous computing power scheduling to generate smooth segmentation boundaries and multi-level rendering picture data; according to the smooth segmentation boundaries, performing picture partition on the multi-level rendering picture data, and outputting a single-frame composite picture stream to a high-definition large-size display terminal.The application solves the problems of the prior art, such as the hard cutting of the line-of-sight when multiple audiences share a large screen, uneven distribution of computing power, and timing jump of the segmentation boundary.

Description

Technical Field

[0001] This invention relates to the field of human-computer interaction technology, and in particular to a method and apparatus for synchronous rendering of screen partitions for appreciating large-scale paintings. Background Technology

[0002] With the rapid development of ultra-high-definition display technology, ultra-high-definition large-format display terminals are increasingly widely used in cultural exhibition venues such as museums and art galleries. In order to provide an immersive experience, existing technologies have introduced eye-tracking and perspective correction rendering technologies (such as naked-eye 3D effects), which calculate the perspective image that conforms to the physical perspective of the viewer in real time based on the position and orientation of the viewer's head.

[0003] However, when multiple viewers simultaneously stand in front of the same giant screen to view the same painting, a contradiction emerges: because each viewer stands in a different position on the screen (left, right, near or far), their corresponding "correct perspective" is completely different (i.e., heterogeneous image). If the screen only displays one perspective, it will inevitably cause severe image distortion and dizziness for viewers in other positions; if the image is output independently for each viewer (such as with a VR headset), the public display significance of "multi-person viewing" on a giant screen is lost.

[0004] The current conventional solution for perspective rendering on multi-user giant screens is "hard screen splitting," which involves simply dividing the screen in half, rendering the perspective for the left side to the left-side viewer, and the perspective for the right side to the right-side viewer. However, this existing technology has the following fatal flaws: 1) Rigid and inflexible dividing lines: They are usually divided by a fixed ratio or a simple average. When the audience's positions intersect or gather on one side, the fixed dividing lines will directly cut off the audience's focus of vision, causing a serious visual break.

[0005] 2) Rigid computing power allocation: Regardless of the number of viewers or the complexity of the scene, the same GPU computing power is allocated to both the left and right sides, resulting in stuttering in areas with dense viewers and many details, while wasting computing power in areas with no viewers.

[0006] 3) Drastic jumps between frames: When the system recalculates the position of the dividing line based on the viewer's movement, the lack of a timing smoothing mechanism causes the dividing boundary to produce high-frequency jitter or instantaneous jumps on the screen, which greatly reduces the viewing experience. Summary of the Invention

[0007] This invention provides a method and apparatus for synchronous rendering of screen partitions for appreciating large-scale paintings. This invention solves the problems of abruptly cutting off the viewpoint by partition lines, uneven distribution of computing power, and temporal jumps at partition boundaries in existing technologies when multiple viewers share a large screen.

[0008] In a first aspect, embodiments of the present invention provide a method for synchronous rendering of image partitions for appreciating large-scale paintings, the method comprising: Using an infrared motion capture system deployed around a high-definition giant display terminal, six degrees of freedom motion data of each viewer's head are collected. After head posture calculation and gaze mapping, a three-dimensional gaze feature set of all viewers on the giant painting is generated. Based on the three-dimensional line-of-sight feature set, the improved PSO algorithm is invoked. Through chaotic sequence initialization, convergence factor improvement and escape repulsion evolution, the optimal screen segmentation and computing power weight scheme for screen partition synchronous rendering is obtained. Based on the optimal screen segmentation and computing power weighting scheme, temporal smoothing and heterogeneous computing power scheduling are performed to generate smooth segmentation boundaries without jumps for the giant painting and multi-level rendering image data containing high and low resolution layers. Based on the smooth segmentation boundary, the multi-level rendered image data is divided into image partitions, and a single-frame composite image stream adapted to all viewers is output to the high-definition giant display terminal.

[0009] The technical solution provided in this application has at least the following beneficial effects: By acquiring six-DOF motion data of the audience through an infrared motion capture system, and accurately calculating the three-dimensional gaze feature set, the screen division lines can move in real time to follow the focus of attention of all viewers, ensuring that high-attention areas are always at the center of high-quality rendering. This is achieved through improvements to particle swarm optimization. The "escape repulsion evolution" in the Optimization (PSO) algorithm defines the high-density view area of ​​the audience as a "dangerous dead zone," forcing particles representing the dividing line to actively avoid the view focus. This achieves an intelligent layout where the dividing line "flexibly avoids" the viewer's position, completely solving the problem of abruptly cutting off the view. It uses Logistic chaotic mapping instead of traditional random initialization, ensuring that the particle swarm uniformly traverses the entire solution space in the initial stage, effectively avoiding the traditional PSO algorithm's tendency to get stuck in local dead zones in complex multi-viewer scenarios. The fitness function innovatively couples three variables—view aggregation cost, artistic semantic cost, and computational time constraint—for joint optimization. It also introduces a semantic gradient mask map of the artwork, ensuring that when computational power is limited, semantically rich areas (such as the faces of figures in the artwork) receive higher computational weight and resolution (heterogeneous scheduling), while edge areas have reduced resolution, ensuring a stable global frame rate. The resolution scaling factor of the left and right rendering engines is dynamically adjusted according to the optimal computational weight ratio, making the most of the limited graphics processing unit (GPU). The computing power of the GPU is precisely directed to the focal area where the audience is densely focused. To address the abrupt changes in the dividing lines caused by group movement, a one-dimensional Kalman filter is introduced. By reasonably calibrating process noise and observation noise, a low-pass filter is implemented for the coordinates of the dividing lines, filtering out high-frequency jitter and presenting the audience with a silky smooth transition boundary. This completely eliminates the inter-frame jump phenomenon of the dividing lines on the giant screen and ensures extremely high visual continuity.

[0010] In one alternative implementation, an infrared motion capture system deployed around a high-definition giant display terminal is used to collect six-DOF motion data of each viewer's head. After head pose calculation and gaze mapping, a three-dimensional gaze feature set for all viewers on the giant painting is generated, including: Using an infrared motion capture system deployed around a high-definition giant display terminal, the system tracks reflective markers on each audience member's headband and uses rigid body calculation algorithms to collect six-degree-of-freedom motion data of each audience member's head in the physical world coordinate system. Based on the pre-calibrated average gaze reference vector of the human eye in the head coordinate system, the head posture is calculated from the head rotation transformation matrix in the six-degree-of-freedom motion data to obtain the three-dimensional gaze direction vector in the world coordinate system. Based on the three-dimensional gaze direction vector, and taking the three-dimensional coordinates of the head centroid in the six-degree-of-freedom motion data as the starting point, the gaze ray equation is constructed. The line-of-sight ray equation is mapped to the screen physical plane equation of the high-definition giant display terminal to calculate the absolute two-dimensional gaze point coordinates of each viewer on the giant painting. By integrating the absolute two-dimensional gaze point coordinates of all viewers and the corresponding head space depth values ​​in the three-dimensional coordinates of the head centroid, a three-dimensional gaze feature set of all viewers on the giant painting is generated.

[0011] In one optional implementation, the six-degree-of-freedom motion data includes the three-dimensional coordinates of the head's center of mass and the head rotation transformation matrix; The three-dimensional coordinates of the head centroid include the two-dimensional coordinates of the viewer's head in the screen physical plane coordinate system and the spatial depth value of the viewer's head. The head rotation transformation matrix includes the yaw angle orthogonal rotation matrix, the pitch angle orthogonal rotation matrix, and the roll angle orthogonal rotation matrix.

[0012] In one optional implementation, based on the 3D view feature set, an improved PSO algorithm is invoked. Through chaotic sequence initialization, convergence factor improvement, and escape repulsion evolution, iterative optimization is performed to obtain the optimal screen segmentation and computational weighting scheme for synchronized rendering of screen partitions, including: The horizontal coordinates of the vertical dividing line of the high-definition large-format display terminal and the computing power weight ratio of the left and right partition GPUs of the back-end server are encoded into the position vectors of particles in the improved PSO algorithm. A fitness function is defined, and the three-dimensional line-of-sight feature set is used as the simulation environment; The chaotic sequence is generated using the Logistic chaotic mapping, and then mapped to the solution space of the particles in the improved PSO algorithm to obtain the initial particle swarm. Based on the simulated environment, the fitness function is used to calculate the fitness value of each initial particle in the initial particle swarm. Based on the fitness value and the pre-set danger coefficient threshold, the optimal position of each initial particle, the global optimal particle in the initial particle swarm, and the dangerous dead zone coordinate set are determined. The Logistic chaotic mapping is used to generate chaotic seeds, and based on the chaotic seeds, the chaotic mutation rate is generated. The dangerous dead zone coordinate set is called to calculate the escape repulsion force rate. Based on the optimal particle position and the global optimal particle, a convergence factor, chaotic mutation rate, and escape repulsion force rate are introduced to update the PSO position of the initial particle swarm or the updated particle swarm from the previous iteration, thus obtaining the updated particle swarm. Repeatedly update the position of the particle swarm. When the current iteration reaches the maximum iteration count or the fitness value of the globally optimal particle meets the requirements, terminate the iterative update of the particle swarm and output the globally optimal particle at the current iteration count. The position vector of the globally optimal particle is decoded to obtain the optimal screen segmentation and computing power weight scheme for the current moment of screen partition synchronous rendering. The optimal screen segmentation and computing power weight scheme for the current moment includes the horizontal coordinate of the optimal vertical segmentation line and the optimal computing power weight ratio for the current moment.

[0013] In one alternative implementation, the fitness function is formulated as follows: In the formula, For the first i Particles fitness value; For the first in the simulation environment j The absolute two-dimensional gaze point coordinates of each viewer; M Total number of viewers; The x-coordinate of the alternative vertical dividing line As a boundary, the first in the simulation environment j Each audience member is assigned to the center coordinates of the area on the left or right; j Indicate the quantity for the audience; For the first i Particles The x-coordinate of the alternative vertical dividing line; For the first i Particles The weighting ratio of alternative computing power; For input The resulting semantic gradient mask of the painting; It is a first-order differential; For alternative computing power weight ratio The rendering time of the alternative solutions; Rendering time for the target solution; For fitness weighting coefficients; i This is a particle indicator.

[0014] In one alternative implementation, the formula for initializing the chaotic sequence is: In the formula, For the first n+ 1. n One chaotic variable; The stability coefficient; In the formula, For the initial particle swarm, the first i An initial particle; For the first i One chaotic variable; To determine the upper and lower bounds of the solution space;i For particle indication; t This represents the current iteration number; In the formula, For the initial particle swarm, the first i The initial velocity of the initial particle; For interval A function for randomly selecting values.

[0015] In one optional implementation, based on an optimal screen segmentation and computing power weighting scheme, temporal smoothing and heterogeneous computing power scheduling are performed to generate smooth, non-abrupt segmentation boundaries for the giant painting and multi-level rendered image data containing high and low resolution layers, including: The optimal screen segmentation and computing power weighting scheme is analyzed to obtain the optimal vertical segmentation line x-coordinate and the optimal computing power weighting ratio at the current time. Based on the x-coordinate of the optimal vertical dividing line at the previous moment, a one-dimensional Kalman filter is used to smooth the x-coordinate of the optimal vertical dividing line at the current moment using a low-pass filter, resulting in a smooth dividing boundary of the giant painting at the current moment without any jumps. Based on the optimal computing power weight ratio at the current moment, the resolution scaling factor of the rendering engine is dynamically adjusted, and the adjusted rendering engine is used to render the left and right regions separately, generating multi-level rendering image data containing high and low resolution layers.

[0016] In one optional implementation, based on the optimal vertical dividing line x-coordinate of the previous time step, a one-dimensional Kalman filter is used to perform low-pass filtering smoothing on the optimal vertical dividing line x-coordinate of the current time step, resulting in a smooth dividing boundary of the giant painting at the current time step without jumps, including: Based on the x-coordinate of the optimal vertical segmentation line at the previous time step, predict the prior position of the smooth segmentation boundary at the current time step. The prior prediction error at the current time is obtained by superimposing the process noise on the posterior error of the previous time step, and then the Kalman gain at the current time step is calculated using a one-dimensional Kalman filter. Using the x-coordinate of the optimal vertical dividing line at the current moment as the observation value, and combining it with the Kalman gain, the prior position of the smooth dividing boundary at the current moment is corrected to obtain the smooth dividing boundary of the giant painting at the current moment without jumps. The posterior error at the current time step is passed to the one-dimensional Kalman filter at the next time step.

[0017] In one optional implementation, the multi-level rendered image data is partitioned according to smooth segmentation boundaries, and a single-frame composite image stream adapted to all viewers is output to a high-definition large-format display terminal, including: Using smooth segmentation boundaries as the dividing line, a binary Alpha mask matrix is ​​generated at the full-screen resolution of a high-definition large-format display terminal. In the independent rendering contexts of the left and right regions, the parsed data UI requested by the audience is overlaid onto the corresponding multi-level rendering screen data layers; Based on the binarized Alpha mask matrix, multi-level rendered image data is synthesized into a single frame composite image stream through pixel-level multiplication and addition operations, and then pushed to the high-definition large-format display terminal.

[0018] Secondly, embodiments of the present invention provide a device for synchronous rendering of image partitions for appreciating large-scale paintings, used to implement a method for synchronous rendering of image partitions for appreciating large-scale paintings. The device includes: The infrared motion capture unit is used to collect six degrees of freedom motion data of each viewer's head using an infrared motion capture system deployed around the high-definition giant display terminal. After head posture calculation and gaze mapping, a three-dimensional gaze feature set of all viewers on the giant painting is generated. The iterative optimization unit is used to call the improved PSO algorithm based on the three-dimensional line-of-sight feature set, and to perform iterative optimization through chaotic sequence initialization, convergence factor improvement and escape repulsion evolution to obtain the optimal screen segmentation and computing power weight scheme for screen partition synchronous rendering. The synchronous rendering unit is used to perform temporal smoothing and heterogeneous computing power scheduling based on the optimal screen segmentation and computing power weight scheme, and to generate smooth segmentation boundaries without jumps for the giant painting and multi-level rendering image data containing high and low resolution layers. The screen partitioning unit is used to partition the multi-level rendered screen data according to the smooth segmentation boundary, and output a single-frame composite screen stream adapted to all viewers to the high-definition giant display terminal.

[0019] A third aspect of this invention provides an electronic device, which includes: At least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by at least one processor, such that the at least one processor can perform the method proposed in the first aspect of the present invention.

[0020] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described in the first aspect of the present invention. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the electronic device structure of the hardware operating environment involved in the embodiments of the present invention; Figure 2This is a flowchart illustrating the steps of a method for synchronous rendering of image partitions for appreciating large-scale paintings, provided in an embodiment of the present invention. Figure 3 This is a schematic diagram of the functional units of a screen partitioning synchronous rendering device for appreciating large-scale paintings provided in an embodiment of the present invention. Detailed Implementation

[0022] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0023] The present invention will be further described below with reference to the accompanying drawings.

[0024] Reference Figure 1 , Figure 1 This is a schematic diagram of the electronic device structure of the hardware operating environment involved in the embodiments of the present invention.

[0025] like Figure 1 As shown, the electronic device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM), such as a disk drive. The memory 1005 may also optionally be a storage device independent of the aforementioned processor 1001.

[0026] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0027] like Figure 1As shown, the memory 1005, which serves as a storage medium, may include an operating system, a data storage module, a network communication module, a user interface module, and an electronic program for a screen partitioning synchronous rendering device for appreciating large-scale paintings.

[0028] exist Figure 1 In the electronic device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the electronic device of the present invention can be set in the electronic device. The electronic device calls the electronic program of the large-scale painting appreciation screen partition synchronous rendering device stored in the memory 1005 through the processor 1001, and executes the large-scale painting appreciation screen partition synchronous rendering method provided in the embodiment of the present invention.

[0029] Reference Figure 2 The present invention provides a method for synchronous rendering of screen partitions for appreciating large-scale paintings, the method comprising: S201: Using an infrared motion capture system deployed around a high-definition giant display terminal, six degrees of freedom motion data of each viewer's head are collected. After head posture calculation and gaze mapping, a three-dimensional gaze feature set of all viewers on the giant painting is generated. S202: Based on the three-dimensional line-of-sight feature set, the improved PSO algorithm is invoked. Through chaotic sequence initialization, convergence factor improvement and escape repulsion evolution, the optimal screen segmentation and computing power weight scheme for screen partition synchronous rendering is obtained. S203: Based on the optimal screen segmentation and computing power weighting scheme, time-series smoothing and heterogeneous computing power scheduling are performed to generate smooth segmentation boundaries without jumps for the giant painting and multi-level rendering image data containing high and low resolution layers. S204: Based on the smooth segmentation boundary, the multi-level rendered image data is divided into image partitions, and a single-frame composite image stream adapted to all viewers is output to the high-definition giant display terminal.

[0030] The technical solution provided in this application has at least the following beneficial effects: By acquiring six-DOF motion data of the audience through an infrared motion capture system, a three-dimensional gaze feature set is accurately calculated. This allows the screen dividing lines to move in real time following the focus of attention of all viewers, ensuring that high-attention areas are always at the center of high-quality rendering. By improving the "escape repulsion evolution" in the PSO algorithm, high-density gaze areas of the audience are defined as "danger dead zones," forcing particles representing the dividing lines to actively avoid the gaze focus. This achieves an intelligent layout where the dividing lines "flexibly avoid" the audience's position, completely solving the problem of abruptly cutting off the gaze. Logistic chaotic mapping is used instead of traditional random initialization, ensuring that the particle swarm uniformly traverses the entire solution space in the initial stage, effectively avoiding the defect of the traditional PSO algorithm that is prone to getting stuck in local dead zones in complex multi-viewer scenarios. The fitness function is innovatively coupled with the view... The algorithm employs a three-variable optimization approach, considering line aggregation cost, artistic semantic cost, and computational time constraints. It also introduces a semantic gradient mask map of the artwork to ensure that, when computational power is limited, semantically rich regions (such as the faces of figures in paintings) receive higher computational weight and resolution (heterogeneous scheduling), while edge regions have reduced resolution to maintain a stable global frame rate. The resolution scaling factor of the left and right rendering engines is dynamically adjusted based on the optimal computational weight ratio, precisely allocating limited GPU computing power to the densely populated and focused areas of the audience. To address the abrupt changes in the dividing lines caused by group movement, a one-dimensional Kalman filter is introduced. By reasonably calibrating process noise and observation noise, a low-pass filter is implemented for the dividing line coordinates, filtering out high-frequency jitter and presenting the audience with silky smooth transition boundaries. This completely eliminates the inter-frame jump phenomenon of dividing lines on a large screen, ensuring extremely high visual continuity.

[0031] In one alternative implementation, the runtime environment of a method for synchronous rendering of screen partitions for appreciating large-scale paintings includes: a front-end perception layer, a central decision layer, and a back-end rendering layer. The front-end perception layer includes an infrared motion capture system consisting of at least four infrared motion capture cameras (such as the OptiTrack series) deployed around a high-definition giant display terminal (such as a 4m×2m, 8K resolution LED screen), as well as lightweight reflective headbands worn by the audience to collect six degrees of freedom motion data of each audience member's head. After head posture calculation and gaze mapping, a three-dimensional gaze feature set of all audience members on the giant painting is generated. The central decision-making layer includes a back-end server equipped with a high-performance CPU, which is used to call the improved PSO algorithm based on the three-dimensional line-of-sight feature set. Through chaotic sequence initialization, convergence factor improvement and escape repulsion evolution, iterative optimization is performed to obtain the optimal screen segmentation and computing power weight scheme for screen partition synchronous rendering. The backend rendering layer includes a heterogeneous GPU cluster (e.g., two NVIDIA RTX 4090-level graphics cards) to perform temporal smoothing and heterogeneous computing power scheduling based on the optimal screen segmentation and computing power weighting scheme, generating smooth, non-abrupt segmentation boundaries for the giant painting and multi-level rendering image data containing high and low resolution layers; according to the smooth segmentation boundaries, the multi-level rendering image data is partitioned into image partitions, and a single-frame composite image stream adapted to all viewers is output to the high-definition giant display terminal.

[0032] In one alternative implementation, an infrared motion capture system deployed around a high-definition giant display terminal is used to collect six-DOF motion data of each viewer's head. After head pose calculation and gaze mapping, a three-dimensional gaze feature set for all viewers on the giant painting is generated, including: S2011: Using an infrared motion capture system deployed around a high-definition giant display terminal, the system tracks reflective markers on each viewer's headband and uses a rigid body calculation algorithm to collect six degrees of freedom motion data of each viewer's head in the physical world coordinate system. S2012: Based on the pre-calibrated average gaze reference vector in the head coordinate system, the head posture is calculated from the head rotation transformation matrix in the six-degree-of-freedom motion data to obtain the three-dimensional gaze direction vector in the world coordinate system. The formula is as follows: In the formula, This is the three-dimensional line-of-sight direction vector in the world coordinate system; This is the reference vector for the average line of sight of the human eye in the head coordinate system; This is the head rotation transformation matrix; It is the transpose symbol; The yaw angle is the standard orthogonal rotation matrix; The pitch angle is a standard orthogonal rotation matrix; The roll angle is the standard orthogonal rotation matrix; These are the yaw angle, pitch angle, and roll angle. S2013: Based on the three-dimensional gaze direction vector, and taking the three-dimensional coordinates of the head centroid in the six-degree-of-freedom motion data as the starting point, construct the gaze ray equation, the formula is: In the formula, The equation for the line-of-sight ray represents the equation that varies with a specific distance parameter. d The coordinates of any moving point on the changing line of sight; d For distance parameter variables; The three-dimensional coordinates of the head's centroid; S2014: Map the line-of-sight ray equation to the screen physical plane equation of the high-definition giant display terminal to calculate the absolute two-dimensional gaze point coordinates of each viewer on the giant painting. Solve the equations of the line-of-sight ray and the physical plane of the screen using analytical geometry. The intersection of, where, The parameters of the plane equations for a high-definition, large-format display terminal in the global physical world coordinate system. These are absolute spatial coordinate variables in the global physical world coordinate system; The formula is: In the formula, These are two-dimensional screen coordinates in the screen pixel coordinate system, forming the absolute two-dimensional gaze point coordinates. ; This represents the components of the 3D line-of-sight direction vector in the world coordinate system. The specific distance parameter at which the line of sight hits the screen; S2015: Integrate the absolute two-dimensional gaze point coordinates of all viewers and the corresponding head spatial depth values ​​in the three-dimensional coordinates of the head centroid to generate a three-dimensional gaze feature set for all viewers on the large-scale painting. The formula is as follows: In the formula, A three-dimensional line-of-sight feature set; For the first Three-dimensional gaze characteristics of an individual viewer; M This refers to the total number of viewers.

[0033] In one optional implementation, the six-degree-of-freedom motion data includes the three-dimensional coordinates of the head's center of mass and the head rotation transformation matrix; The three-dimensional coordinates of the head centroid Includes the two-dimensional coordinates of the viewer's head in the screen's physical plane coordinate system. and the depth of the audience's head space ; The head rotation transformation matrix includes the yaw angle orthogonal rotation matrix, the pitch angle orthogonal rotation matrix, and the roll angle orthogonal rotation matrix.

[0034] In one optional implementation, based on the 3D view feature set, an improved PSO algorithm is invoked. Through chaotic sequence initialization, convergence factor improvement, and escape repulsion evolution, iterative optimization is performed to obtain the optimal screen segmentation and computational weighting scheme for synchronized rendering of screen partitions, including: S2021: Encode the horizontal coordinate of the vertical dividing line of the high-definition large-format display terminal and the computing power weight ratio of the left and right partition GPUs of the back-end server into the position vector of the particles in the improved PSO algorithm. S2022: Set the fitness function and use the three-dimensional line-of-sight feature set as the simulation environment; S2023: Use Logistic chaotic mapping to generate chaotic sequences, and map the chaotic sequences to the solution space of particles in the improved PSO algorithm to obtain the initial particle swarm; S2024: Based on the simulation environment, using the fitness function, calculate the fitness value of each initial particle in the initial particle swarm, and determine the optimal position of each initial particle, the global optimal particle in the initial particle swarm, and generate a set of dangerous dead zone coordinates based on the fitness value and a pre-set danger coefficient threshold. The formula is: In the formula, The coordinate set of the danger dead zone; The first one that was classified as a dangerous dead zone i A dangerous particle; This is the threshold for the risk factor; For the first t The iteration of the ... i Dangerous particles The fitness value, in the first iteration, For the initial particles; For the first t The worst global particle in the next iteration fitness value; Indicator of hazardous particles; S2025: Use the Logistic chaotic mapping to generate a chaotic seed, and based on the chaotic seed, generate the chaotic mutation rate. Then, call the dangerous dead zone coordinate set to calculate the escape repulsion force velocity. The formula is: In the formula, This refers to the rate of chaotic mutation. Seed of chaos; The velocity of the repulsive force for escape; For escape repulsion force parameters; The total number of particles in the coordinate set of the danger dead zone; S2026: Based on the optimal particle position and the global optimal particle, a convergence factor, chaotic mutation rate, and escape repulsion force rate are introduced to update the PSO position of the initial particle swarm or the updated particle swarm of the previous iteration, so as to obtain the updated particle swarm. In the formula, For the first t+ 1st iteration i The rate at which each new particle is updated; For the firstt The iteration of the ... i The rate at which each new particle is updated, in the first iteration, The initial velocity; For the first t The iteration of the ... i The optimal position of each particle; Accelerate one's own cognition; The acceleration coefficient of social cognition; Let be a random perturbation vector in the interval [0,1]. For the first t The globally optimal particle in the next iteration; For the first t The convergence factor of the next iteration; This refers to the rate of chaotic mutation. The velocity of the repulsive force for escape; In the formula, For the first t+ 1st iteration i An updated particle; For the first t The second iteration i A newer particle, in the first iteration, For the initial particles; S2027: Repeatedly update the position of the particle swarm. When the current iteration number reaches the maximum iteration number or the fitness value of the global best particle meets the requirements, terminate the iterative update of the particle swarm and output the global best particle for the current iteration number. S2028: Decode the position vector of the globally optimal particle to obtain the optimal screen segmentation and computing power weighting scheme for the current moment in the synchronous rendering of screen partitions. The optimal screen segmentation and computing power weighting scheme at the current moment includes the optimal vertical segmentation line x-coordinate at the current moment. And the optimal computing power weight ratio at the current moment ; l This refers to the current moment.

[0035] In one alternative implementation, the fitness function is formulated as follows: In the formula, For the first i Particles fitness value; For the first in the simulation environment j The absolute two-dimensional gaze point coordinates of each viewer; M Total number of viewers; The x-coordinate of the alternative vertical dividing line As a boundary, the first in the simulation environment j Each audience member is assigned to the center coordinates of the area on the left or right; j Indicate the quantity for the audience; For the first i Particles The x-coordinate of the alternative vertical dividing line; For the first i Particles The weighting ratio of alternative computing power; For input The resulting semantic gradient mask of the painting; It is a first-order differential; For alternative computing power weight ratio The rendering time of the alternative solutions; Rendering time for the target solution; For fitness weighting coefficients; i For particle indication; The forced dividing line moves closer to the edge of the densely populated area, so that most viewers’ eyes fall on the center of the high-computing-power area where they are located, thus achieving the best distortion control and clarity experience. In the item This is a semantic gradient mask extracted after performing Sobel or Canny edge detection on the currently displayed large-scale painting. The mask is then analyzed along the vertical direction (y-axis). The cost of performing first-order differentiation and integration increases dramatically if the dividing line cuts through areas with complex textures in the artwork (such as the outlines of figures or the edges of buildings). Therefore, the algorithm actively seeks "flat" areas in the artwork (such as solid-color backgrounds or large areas of sky) as the points through which the dividing line passes. This is the most significant artistic consideration that distinguishes it from ordinary computer desktop splicing technology. Ensure that the allocated weight ratio can be rendered within the target frame time to prevent computing power overload.

[0036] In one alternative implementation, the formula for initializing the chaotic sequence is: In the formula, For the first n+ 1. n There are several chaotic variables whose values ​​range from [0, 1]. The stability coefficient is typically 4. This sequence is ergodic and random, ensuring that the initial particle swarm is uniformly distributed in the solution space, avoiding getting trapped in local optima, which is superior to traditional random initialization. n Indicator of chaotic variables; In the formula, For the initial particle swarm, the first i An initial particle; For the first i One chaotic variable; To determine the upper and lower bounds of the solution space; i For particle indication; t This represents the current iteration number; In the formula, For the initial particle swarm, the first i The initial velocity of the initial particle; For interval A function for randomly selecting values.

[0037] In one optional implementation, based on an optimal screen segmentation and computing power weighting scheme, temporal smoothing and heterogeneous computing power scheduling are performed to generate smooth, non-abrupt segmentation boundaries for the giant painting and multi-level rendered image data containing high and low resolution layers, including: S2031: Analyze the optimal screen segmentation and computing power weighting scheme to obtain the optimal vertical segmentation line x-coordinate and the optimal computing power weighting ratio at the current moment. S2032: Based on the optimal vertical dividing line x-coordinate of the previous time step, use a one-dimensional Kalman filter to perform low-pass filtering smoothing on the optimal vertical dividing line x-coordinate of the current time step, and obtain the smooth dividing boundary of the giant painting at the current time step without jumps. S2033: Based on the optimal computing power weight ratio at the current moment, dynamically adjust the resolution scaling factor of the rendering engine, and use the adjusted rendering engine to render the left and right regions separately, generating multi-level rendering image data containing high and low resolution layers. The formula is: In the formula, v This is the resolution scaling factor adjusted by the rendering engine for the left and right areas of the high-definition large-format display terminal. This represents the optimal computing power weight ratio at the current moment. The optimal computing power weight coefficients for the left and right regions in the optimal computing power weight ratio; For example, =1.0 (i.e., native resolution rendering). =0.65 (using the reverse operation of super-resolution techniques such as spatial downsampling or DLSS / FSR, rendering at a lower resolution and then upscaling), the left and right GPUs perform parallel calculations, ultimately generating a high-definition layer on the left and a low-definition layer on the right, greatly freeing up the video memory and computing bandwidth of the right GPU.

[0038] In one optional implementation, based on the optimal vertical dividing line x-coordinate of the previous time step, a one-dimensional Kalman filter is used to perform low-pass filtering smoothing on the optimal vertical dividing line x-coordinate of the current time step, resulting in a smooth dividing boundary of the giant painting at the current time step without jumps, including: S20321: Based on the x-coordinate of the optimal vertical segmentation line at the previous time step, predict the prior position of the smooth segmentation boundary at the current time step, using the following formula: In the formula, For the current moment l The prior position of the smooth segmentation boundary; For the previous moment l- The optimal vertical dividing line x-coordinate of 1; l The current moment; S20322: The posterior error from the previous time step is superimposed with process noise to obtain the prior prediction error at the current time step. Then, a one-dimensional Kalman filter is used to calculate the Kalman gain at the current time step, as shown in the formula: In the formula, For the current moment l The prior prediction error; For the previous moment l- The posterior error of 1; This is process noise; This is the state transition matrix; T It is the transpose symbol; The Kalman gain at the current moment; To observe noise; l The current moment; S20323: Using the optimal vertical dividing line x-coordinate at the current moment as the observation value, and combining it with the Kalman gain, the prior position of the smooth dividing boundary at the current moment is corrected to obtain the smooth dividing boundary of the giant painting at the current moment without jumps. The formula is: In the formula, For the current moment l Smooth segmentation boundaries; The x-coordinate of the optimal vertical dividing line at the current moment; S20324: The one-dimensional Kalman filter that propagates the posterior error of the current time step to the next time step, is defined by the following formula: In the formula, For the current moment l The posterior error.

[0039] In one optional implementation, the multi-level rendered image data is partitioned according to smooth segmentation boundaries, and a single-frame composite image stream adapted to all viewers is output to a high-definition large-format display terminal, including: S2041: Using the smooth segmentation boundary as the dividing line, generate a binary alpha mask matrix at the full-screen resolution of a high-definition large-format display terminal. The formula is as follows: In the formula, This is the binarized alpha mask matrix for the left and right areas of a high-definition large-format display terminal; For the current moment l Smooth segmentation boundaries; For high-definition large-format display terminals, the two-dimensional screen coordinate variables of the pixels in the screen pixel coordinate system; S2042: In the independent rendering contexts of the left and right regions, the parsed data UI requested by the audience is overlaid onto the corresponding multi-level rendering screen data layers; In this embodiment, in the rendering contexts of the left and right GPUs (such as independent OpenGL / Vulkan DeviceContexts), the low-resolution multi-level rendering image data is upsampled to the full-screen resolution (if downsampling rendering is used). Then, the parsed data UI triggered by the viewer through a tablet or touch screen (for example, a viewer clicks on the upper left corner of the painting to request the display of the creative background text box in that area) is accurately superimposed onto the coordinate system corresponding to the left layer. This distributed UI superposition avoids the additional overhead caused by global UI rendering. S2043: Based on the binarized alpha mask matrix, multi-level rendered image data is synthesized into a single-frame composite image stream through pixel-level multiplication and addition operations, and then streamed to the high-definition large-format display terminal. The formula is as follows: In the formula, This is a single-frame composite video stream; This refers to multi-level rendered image data for the left and right areas of a high-definition large-format display terminal. In this embodiment, since the binarized Alpha mask matrix of the left and right areas of the high-definition large-format display terminal is an absolute binary mask generated based on smooth boundaries, each pixel is only taken from the left or right side, and there is no edge blurring caused by semi-transparent mixing. Due to the temporal smoothness of the boundary, even if there are extremely small differences in brightness / sharpness between the left and right sides due to different resolution scaling factors, this difference is restricted to a smoothly moving line on the spatial axis and is low-pass filtered on the temporal axis, so the human visual system cannot detect the existence of the seam at all.

[0040] This invention also provides a screen partitioning synchronous rendering device 300 for appreciating large-scale paintings, see reference. Figure 3 The device may include the following units: The infrared motion capture unit 301 is used to collect six degrees of freedom motion data of each viewer's head using an infrared motion capture system deployed around the high-definition giant display terminal. After head posture calculation and gaze mapping, a three-dimensional gaze feature set of all viewers on the giant painting is generated. The iterative optimization unit 302 is used to call the improved PSO algorithm based on the three-dimensional line-of-sight feature set, and to perform iterative optimization through chaotic sequence initialization, convergence factor improvement and escape repulsion evolution to obtain the optimal screen segmentation and computing power weight scheme for screen partition synchronous rendering. Synchronous rendering unit 303 is used to perform temporal smoothing and heterogeneous computing power scheduling based on the optimal screen segmentation and computing power weight scheme, and generate smooth segmentation boundaries without jumps for the giant painting and multi-level rendering screen data containing high and low resolution layers. The screen partitioning unit 304 is used to partition the multi-level rendered screen data according to the smooth segmentation boundary, and output a single-frame composite screen stream adapted to all viewers to the high-definition giant display terminal.

[0041] Based on the same inventive concept, another embodiment of the present invention provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus. Memory, used to store computer programs; When a processor executes a program stored in memory, it implements the method for synchronous rendering of screen partitions for appreciating large-scale paintings according to the present invention.

[0042] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EI) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of representation, only one thick line is used in the diagram, but this does not indicate that there is only one bus or one type of bus. The communication interface is used for communication between the aforementioned terminal and other devices. The memory can include Random Access Memory (RAM), or non-volatile memory, such as at least one disk storage device. Optionally, the memory can also be at least one storage device located remotely from the aforementioned processor.

[0043] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0044] Furthermore, to achieve the above objectives, embodiments of the present invention also propose a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for synchronous rendering of screen partitions for appreciating large-scale paintings according to embodiments of the present invention.

[0045] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, embodiments of the present invention can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of computer program products implemented on one or more computer-usable hardware devices (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0046] The embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (apparatus), and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0047] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0048] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0049] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. "And / or" indicates that either one or both can be chosen. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes the element.

[0050] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for synchronous rendering of image zones for appreciating large-scale paintings, characterized in that, The method includes: Using an infrared motion capture system deployed around a high-definition giant display terminal, six degrees of freedom motion data of each viewer's head are collected. After head posture calculation and gaze mapping, a three-dimensional gaze feature set of all viewers on the giant painting is generated. Based on the three-dimensional line-of-sight feature set, the improved PSO algorithm is invoked. Through chaotic sequence initialization, convergence factor improvement and escape repulsion evolution, the optimal screen segmentation and computing power weight scheme for screen partition synchronous rendering is obtained. Based on the optimal screen segmentation and computing power weighting scheme, temporal smoothing and heterogeneous computing power scheduling are performed to generate smooth segmentation boundaries without jumps for the giant painting and multi-level rendering image data containing high and low resolution layers. Based on the smooth segmentation boundary, the multi-level rendered image data is divided into image partitions, and a single-frame composite image stream adapted to all viewers is output to the high-definition giant display terminal.

2. The method for synchronous rendering of image zones for appreciating large-scale paintings according to claim 1, characterized in that, Using an infrared motion capture system deployed around a high-definition giant display terminal, six-degree-of-freedom motion data of each viewer's head is collected. After head posture calculation and gaze mapping, a three-dimensional gaze feature set of all viewers on the giant painting is generated, including: Using an infrared motion capture system deployed around a high-definition giant display terminal, the system tracks reflective markers on each audience member's headband and uses rigid body calculation algorithms to collect six degrees of freedom motion data of each audience member's head in the physical world coordinate system. Based on the pre-calibrated average gaze reference vector of the human eye in the head coordinate system, the head posture is calculated from the head rotation transformation matrix in the six-degree-of-freedom motion data to obtain the three-dimensional gaze direction vector in the world coordinate system. Based on the three-dimensional gaze direction vector, and taking the three-dimensional coordinates of the head centroid in the six-degree-of-freedom motion data as the starting point, the gaze ray equation is constructed. The line-of-sight ray equation is mapped to the screen physical plane equation of the high-definition giant display terminal to calculate the absolute two-dimensional gaze point coordinates of each viewer on the giant painting. By integrating the absolute two-dimensional gaze point coordinates of all viewers and the corresponding head space depth values ​​in the three-dimensional coordinates of the head centroid, a three-dimensional gaze feature set of all viewers on the giant painting is generated.

3. The method for synchronous rendering of image zones for appreciating large-scale paintings according to claim 2, characterized in that, The six-degree-of-freedom motion data includes the three-dimensional coordinates of the head's center of mass and the head rotation transformation matrix; The three-dimensional coordinates of the head centroid include the two-dimensional coordinates of the viewer's head in the screen's physical plane coordinate system and the spatial depth value of the viewer's head. The head rotation transformation matrix includes the yaw angle orthogonal rotation matrix, the pitch angle orthogonal rotation matrix, and the roll angle orthogonal rotation matrix.

4. The method for synchronous rendering of image zones for appreciating large-scale paintings according to claim 3, characterized in that, Based on the 3D view feature set, the improved PSO algorithm is invoked. Through chaotic sequence initialization, convergence factor improvement, and escape repulsion evolution, iterative optimization is performed to obtain the optimal screen segmentation and computing power weighting scheme for simultaneous rendering of screen partitions, including: The horizontal coordinates of the vertical dividing line of the high-definition large-format display terminal and the computing power weight ratio of the left and right partition GPUs of the back-end server are encoded into the position vectors of particles in the improved PSO algorithm. A fitness function is defined, and the three-dimensional line-of-sight feature set is used as the simulation environment; The chaotic sequence is generated using the Logistic chaotic mapping, and then mapped to the solution space of the particles in the improved PSO algorithm to obtain the initial particle swarm. Based on the simulated environment, the fitness function is used to calculate the fitness value of each initial particle in the initial particle swarm. Based on the fitness value and the pre-set danger coefficient threshold, the optimal position of each initial particle, the global optimal particle in the initial particle swarm, and the dangerous dead zone coordinate set are determined. The Logistic chaotic mapping is used to generate chaotic seeds, and based on the chaotic seeds, the chaotic mutation rate is generated. The dangerous dead zone coordinate set is called to calculate the escape repulsion force rate. Based on the optimal particle position and the global optimal particle, a convergence factor, chaotic mutation rate, and escape repulsion force rate are introduced to update the PSO position of the initial particle swarm or the updated particle swarm from the previous iteration, thus obtaining the updated particle swarm. Repeatedly update the position of the particle swarm. When the current iteration reaches the maximum iteration count or the fitness value of the globally optimal particle meets the requirements, terminate the iterative update of the particle swarm and output the globally optimal particle at the current iteration count. The position vector of the globally optimal particle is decoded to obtain the optimal screen segmentation and computing power weight scheme for the current moment of screen partition synchronous rendering. The optimal screen segmentation and computing power weight scheme for the current moment includes the horizontal coordinate of the optimal vertical segmentation line and the optimal computing power weight ratio for the current moment.

5. The method for synchronous rendering of image zones for appreciating large-scale paintings according to claim 4, characterized in that, The formula for the fitness function is: In the formula, For the first i Particles fitness value; For the first in the simulation environment j The absolute two-dimensional gaze point coordinates of each viewer; M Total number of viewers; The x-coordinate of the alternative vertical dividing line As a boundary, the first in the simulation environment j Each audience member is assigned to the center coordinates of the area on the left or right; j Indicate the quantity for the audience; For the first i Particles The x-coordinate of the alternative vertical dividing line; For the first i Particles The weighting ratio of alternative computing power; For input The resulting semantic gradient mask of the painting; It is a first-order differential; For alternative computing power weight ratio The rendering time of the alternative solutions; Rendering time for the target solution; For fitness weighting coefficients; i This is a particle indicator.

6. The method for synchronous rendering of image zones for appreciating large-scale paintings according to claim 5, characterized in that, The formula for initializing the chaotic sequence is: In the formula, For the first n+ 1. n One chaotic variable; The stability coefficient; In the formula, For the initial particle swarm, the first i An initial particle; For the first i One chaotic variable; To determine the upper and lower bounds of the solution space; i For particle indication; t This represents the current iteration number; In the formula, For the initial particle swarm, the first i The initial velocity of the initial particle; For interval A function for randomly selecting values.

7. The method for synchronous rendering of image zones for appreciating large-scale paintings according to claim 6, characterized in that, Based on the optimal screen segmentation and computing power weighting scheme, temporal smoothing and heterogeneous computing power scheduling are performed to generate smooth, seamless segmentation boundaries for the giant painting and multi-level rendered image data containing high and low resolution layers, including: The optimal screen segmentation and computing power weighting scheme is analyzed to obtain the optimal vertical segmentation line x-coordinate and the optimal computing power weighting ratio at the current time. Based on the x-coordinate of the optimal vertical dividing line at the previous moment, a one-dimensional Kalman filter is used to smooth the x-coordinate of the optimal vertical dividing line at the current moment using a low-pass filter, resulting in a smooth dividing boundary of the giant painting at the current moment without any jumps. Based on the optimal computing power weight ratio at the current moment, the resolution scaling factor of the rendering engine is dynamically adjusted, and the adjusted rendering engine is used to render the left and right regions separately, generating multi-level rendering image data containing high and low resolution layers.

8. The method for synchronous rendering of image zones for appreciating large-scale paintings according to claim 7, characterized in that, Based on the optimal vertical dividing line x-coordinate from the previous time step, a one-dimensional Kalman filter is used to smooth the optimal vertical dividing line x-coordinate at the current time step using a low-pass filter, resulting in a smooth, non-abrupt dividing boundary for the entire painting at the current time step, including: Based on the x-coordinate of the optimal vertical segmentation line at the previous time step, predict the prior position of the smooth segmentation boundary at the current time step. The prior prediction error at the current time is obtained by superimposing the process noise on the posterior error of the previous time step, and then the Kalman gain at the current time step is calculated using a one-dimensional Kalman filter. Using the x-coordinate of the optimal vertical dividing line at the current moment as the observation value, and combining it with the Kalman gain, the prior position of the smooth dividing boundary at the current moment is corrected to obtain the smooth dividing boundary of the giant painting at the current moment without jumps. The posterior error at the current time step is passed to the one-dimensional Kalman filter at the next time step.

9. The method for synchronous rendering of image zones for appreciating large-scale paintings according to claim 8, characterized in that, Based on the smooth segmentation boundaries, the multi-level rendered image data is partitioned into image segments, and a single-frame composite image stream adapted to all viewers is output to the high-definition giant display terminal, including: Using smooth segmentation boundaries as the dividing line, a binary Alpha mask matrix is ​​generated at the full-screen resolution of a high-definition large-format display terminal. In the independent rendering contexts of the left and right regions, the parsed data UI requested by the audience is overlaid onto the corresponding multi-level rendering screen data layers; Based on the binarized Alpha mask matrix, multi-level rendered image data is synthesized into a single frame composite image stream through pixel-level multiplication and addition operations, and then pushed to the high-definition large-format display terminal.

10. A device for synchronous rendering of image partitions for appreciating large-scale paintings, used to implement the synchronous rendering method for image partitions for appreciating large-scale paintings as described in any one of claims 1-9, characterized in that... The device includes: The infrared motion capture unit is used to collect six degrees of freedom motion data of each viewer's head using an infrared motion capture system deployed around the high-definition giant display terminal. After head posture calculation and gaze mapping, a three-dimensional gaze feature set of all viewers on the giant painting is generated. The iterative optimization unit is used to call the improved PSO algorithm based on the three-dimensional line-of-sight feature set, and to perform iterative optimization through chaotic sequence initialization, convergence factor improvement and escape repulsion evolution to obtain the optimal screen segmentation and computing power weight scheme for screen partition synchronous rendering. The synchronous rendering unit is used to perform temporal smoothing and heterogeneous computing power scheduling based on the optimal screen segmentation and computing power weight scheme, and to generate smooth segmentation boundaries without jumps for the giant painting and multi-level rendering image data containing high and low resolution layers. The screen partitioning unit is used to partition the multi-level rendered screen data according to the smooth segmentation boundary, and output a single-frame composite screen stream adapted to all viewers to the high-definition giant display terminal.

Images