Display uniformity guarantee method and system for high-credibility man-machine interaction of robot

Abstract

The invention relates to the technical field of man-machine interaction, and discloses a display uniformity guarantee method and system for high-credibility man-machine interaction of a robot, and the method comprises the steps: fusing real-time output image data and ambient light spectrum data of a display region in a robot display device to obtain multi-dimensional display state information, evaluating independent supplementary demand values of pixels in the display area; performing brightness conduction correlation on the pixels to obtain visual correlation constraints; performing equalization adjustment on the independent supplement demand value to obtain an initial compensation parameter; generating visual saliency weight distribution around the current fixation point of the user, and combining the visual saliency weight distribution with the visual saliency weight distribution to obtain a perception optimization parameter; generating an enhanced display picture based on the perception optimization parameter, and evaluating a quality quantitative index of the enhanced display picture; dynamically adjusting the parameter weight of the equalization adjustment according to a quality quantification index; according to the invention, the efficiency of guaranteeing the display uniformity of high-credibility man-machine interaction of the robot can be improved.

Description

Technical Field

[0001] This invention relates to the field of human-computer interaction technology, and in particular to a method and system for ensuring display uniformity in high-reliability human-computer interaction for robots. Background Technology

[0002] In high-reliability human-computer interaction scenarios involving robots, the uniformity of the display device directly affects the accuracy of interactive information transmission and user experience. Current technologies, while ensuring display uniformity, fail to fully integrate key data from both the display device's output and the environment. They rely solely on single real-time output image data for analysis, neglecting the impact of key parameters such as light intensity and color temperature in the ambient light spectrum on the display state. This results in an inability to fully acquire multi-dimensional display state information of the display area, leading to biased assessments of individual pixel compensation requirements and making it difficult to accurately determine the compensation needs of each pixel. This fundamentally limits the effectiveness of ensuring display uniformity.

[0003] Existing technologies have significant shortcomings in the generation and optimization of compensation parameters. On the one hand, they fail to establish brightness transmission relationships between pixels based on the spatial distribution characteristics of independent pixel supplementation requirements, lacking effective visual correlation constraints to regulate the equalization adjustment process. This makes it easy for initial compensation parameters to exhibit local overcompensation or undercompensation. On the other hand, they do not consider user visual perception patterns, fail to construct a visual saliency weight distribution based on the user's current gaze point to optimize compensation parameters, and lack a display quality quantification evaluation mechanism based on user pupil response data. Furthermore, they cannot dynamically adjust compensation parameters according to actual user perception feedback, resulting in display uniformity failing to adapt to real-time user perception needs and failing to meet the continuous display quality requirements of high-reliability human-computer interaction. Therefore, how to achieve accurate and dynamic assurance of display uniformity through multi-source data fusion, visual correlation constraint construction, and user perception feedback linkage has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a method and system for ensuring display uniformity in high-reliability human-computer interaction for robots, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a method for ensuring display uniformity in high-reliability human-computer interaction for robots, comprising: A method for ensuring display uniformity in high-reliability human-computer interaction for robots, characterized in that the method includes: S1. By fusing real-time output image data and ambient light spectrum data of the display area in the robot display device, multi-dimensional display status information of the display area is obtained, and by combining the multi-dimensional display status information, the independent supplementation requirement value of the pixels in the display area is evaluated. S2. Based on the spatial distribution characteristics of the independent supplementary demand values, perform brightness transmission association on the pixels to obtain the visual association constraints of the pixels; S3. Using the visual association constraint as a limiting condition, the independent supplementary demand value is adjusted to obtain the initial compensation parameters of the display area. S4. Using the user's current gaze point as the center, smoothly decrease the weights outwards to generate the visual saliency weight distribution of the current gaze point, and combine the visual saliency weight distribution with the initial compensation parameters to obtain the perceptual optimization parameters of the display area. S5. Input the perception optimization parameters into the display device to obtain the enhanced display image of the display area, and evaluate the display quality of the enhanced display image based on the user's pupil response data to obtain the quality quantification index of the enhanced display image; S6. Based on the quality quantification index, dynamically adjust the parameter weights of the equalization adjustment.

[0006] In a preferred embodiment, the real-time output image data and ambient light spectrum data of the display area in the fusion robot display device are used to obtain multi-dimensional display state information of the display area. Based on this multi-dimensional display state information, the independent supplementation requirement value of pixels within the display area is evaluated, including: Collect real-time output image data and ambient light spectrum data from the display area of ​​the robot's display device; Photometric measurement processing is performed on the real-time output image data to obtain the brightness distribution characteristics of the real-time output image data; The ambient light characteristics of the ambient light spectrum data are obtained by analyzing the light intensity and color temperature parameters of the ambient light spectrum data. In the spatial and temporal dimensions, the brightness distribution features and the ambient light features are fused to obtain multidimensional display state information of the display area; Based on the multidimensional display status information, the brightness value of the pixel in the display area is compared with the preset interval reference brightness to obtain the brightness deviation data of the pixel. Based on the brightness deviation, the required brightness compensation value for the pixel is derived, and the brightness compensation value is used as the independent supplementary requirement value for the pixel.

[0007] In a preferred embodiment, the step of performing luminance conduction correlation on the pixels based on the spatial distribution characteristics of the independent supplementary demand values ​​to obtain the visual correlation constraints of the pixels includes: Based on the spatial distribution characteristics of the independent supplementary demand values, the boundary information of the high-demand area and the low-demand area in the display area is obtained; Based on the boundary information, spatial continuity analysis is performed on adjacent pixels to obtain the brightness transmission relationship of the adjacent pixels; Based on the brightness transmission relationship, a visual association graph of the pixels is constructed, with the pixels as nodes, the adjacency relationships of the pixels as edges, and the brightness transmission intensity of the adjacent pixels as edge weights. Semantic extraction is performed on the visual association map to obtain the visual association constraints of the pixels.

[0008] In a preferred embodiment, the step of equalizing the independent supplementary demand value based on the visual association constraint to obtain the initial compensation parameters for the display area includes: The visual association constraint is mapped to the compensation transmission rule of the adjacent pixels; According to the compensation conduction rule, the pixel is used as a node, the adjacency relationship of the pixel is used as a conduction link, and the conduction strength in the compensation conduction rule is used as the conduction link weight to construct the compensation conduction network of the display area. The independent supplementary demand value is used as the transmission initiation amount and transmitted along the compensation transmission network to the surrounding area of ​​adjacent pixels, while receiving compensation feedback from the surrounding area. Based on the compensation feedback, the real-time compensation amount of the pixel is adjusted, and the difference stability evaluation of the real-time compensation amount is performed to obtain the stability index of the real-time compensation amount. When the stability index reaches the preset stability condition, the real-time compensation amount is used as the initial compensation parameter for the display area.

[0009] In a preferred embodiment, adjusting the real-time compensation amount of the pixel based on the compensation feedback and performing a difference stability evaluation on the real-time compensation amount to obtain a stability index of the real-time compensation amount includes: The difference values ​​of the real-time compensation amounts of the adjacent pixels are statistically analyzed to obtain the compensation difference set of the adjacent pixels; The distribution pattern of the compensation difference set is analyzed to obtain the global difference level of the compensation difference set; The global difference levels are rearranged in chronological order to obtain a time sequence of the global difference levels. The time series is evaluated for trend development to obtain a stable index of the real-time compensation amount.

[0010] In a preferred embodiment, the step of smoothly decreasing the weights outwards from the user's current gaze point to generate a visual saliency weight distribution for the current gaze point includes: Centered on the user's current gaze point, the display area is divided into a central gaze area, an intermediate transition area, and an outer perception area; Determine the adjacent boundary regions of the central gaze area, the intermediate transition area, and the peripheral sensing area; Based on the distance from the current gaze point to the adjacent boundary region, distance-weighted interpolation is performed on the regional baseline weights of the central gaze region, the intermediate transition region, and the peripheral perception region to obtain the boundary transition weights of the adjacent boundary regions. By integrating the regional baseline weights and the boundary transition weights, the visual saliency weight distribution of the current gaze point is obtained.

[0011] In a preferred embodiment, combining the visual saliency weight distribution and the initial compensation parameters to obtain the perceptual optimization parameters of the display area includes: The pixel weight values ​​in the visual saliency weight distribution are weighted and fused with the corresponding pixel compensation values ​​in the initial compensation parameters to obtain the weighted compensation parameters for the pixel. The calculation formula for the weighted fusion is as follows: ; In the formula, Indicates the first Weighted compensation parameters for each pixel. This represents the first element in the visual saliency weight distribution. Visual saliency weights for each pixel. This represents the first element in the visual saliency weight distribution. Visual saliency weights for each pixel. Indicates the first Initial compensation parameters for each pixel. This represents the preset weight adjustment factor. This represents the total number of pixels within the display area. This represents the mean of the visual saliency weight distribution. The standard deviation of the visual saliency weight distribution is represented by the following: This represents the preset normalization adjustment coefficient. This represents the square root operation. This represents the summation operation; Based on the spatial gradient characteristics of the visual saliency weight distribution, the weighted compensation parameters are spatially consistent to obtain the gradient coordination parameters of the weighted compensation parameters. Based on the regional characteristics of the current gaze point, the gradient coordination parameters are fused according to regional characteristics to obtain the perceptual optimization parameters of the display area.

[0012] In a preferred embodiment, the step of evaluating the display quality of the enhanced display screen based on the user's pupil response data to obtain a quantitative quality index of the enhanced display screen includes: Collect the user's pupil diameter change data; By performing dual-source data coupling on the pupil oscillation frequency and amplitude changes in the pupil diameter change data, the comprehensive stress deviation of the pupil diameter change data is obtained; The overall stress deviation is mapped to the enhanced display screen to obtain the visual stress region distribution of the enhanced display screen; Dwell point density analysis is performed on the movement trajectory of the current gaze point to obtain the visual attention distribution of the enhanced display screen; Spatial overlap calibration is performed on the distribution of visual stress area and visual attention distribution to obtain the high response superposition area of ​​the enhanced display image; The quality quantification index of the enhanced display image is generated based on the area ratio of the high-response overlay region and the average stress intensity.

[0013] In a preferred embodiment, dynamically adjusting the parameter weights of the equalization adjustment based on the quality quantification index includes: The quality quantification index is compared with the preset quality benchmark value to obtain the quality deviation parameter of the enhanced display screen. Based on the quality deviation parameter, confirm the weight adjustment amount of the visual association constraint and the independent supplementary demand value in the equalization adjustment; Based on the weight adjustment amount, the parameter weight configuration of the equalization adjustment is dynamically updated.

[0014] To address the aforementioned problems, the present invention also provides a display uniformity assurance system for high-reliability human-computer interaction in robots, the system comprising: The display status information acquisition module is used to fuse real-time output image data and ambient light spectrum data of the display area in the robot display device to obtain multi-dimensional display status information of the display area, and to evaluate the independent supplementation requirement value of the pixels in the display area in combination with the multi-dimensional display status information. The visual association constraint determination module is used to perform brightness transmission association on the pixel based on the spatial distribution characteristics of the independent supplementary demand value, so as to obtain the visual association constraint of the pixel. The initial compensation parameter generation module is used to equalize and adjust the independent supplementary demand value based on the visual association constraint to obtain the initial compensation parameters of the display area. The visual saliency weight generation module is used to smoothly decrease the weights outward from the user's current gaze point to generate the visual saliency weight distribution of the current gaze point, and combine the visual saliency weight distribution and the initial compensation parameters to obtain the perceptual optimization parameters of the display area. The display quality assessment module is used to input the perception optimization parameters into the display device to obtain the enhanced display image of the display area, and to evaluate the display quality of the enhanced display image based on the user's pupil response data to obtain the quality quantification index of the enhanced display image. The parameter weight adjustment module is used to dynamically adjust the parameter weights of the equalization adjustment based on the quality quantification index.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention, by fusing real-time output image data from a robot display device with ambient light spectrum data, can comprehensively acquire multi-dimensional display state information of the display area. Based on this, it can accurately assess the independent supplementation demand values ​​of pixels within the display area, providing precise data support for ensuring display uniformity. Simultaneously, based on the spatial distribution characteristics of the independent supplementation demand values, it constructs brightness transmission relationships between pixels, forming visual association constraints and standardizing the equalization adjustment process. The generated initial compensation parameters can achieve pixel-level equalization compensation, effectively improving the basic accuracy of display uniformity assurance and enhancing the accuracy of interactive information transmission.

[0016] 2. This invention combines user visual perception patterns with a visual saliency weight distribution centered on the user's current gaze point. This distribution, combined with initial compensation parameters, yields perception optimization parameters, enabling display uniformity to adapt to the user's visual focus areas and improving display quality at the user's perception level. Furthermore, by generating quality quantification indicators using user pupil response data and dynamically adjusting the weights of the equalization adjustment parameters, dynamic optimization of display uniformity can be achieved, continuously matching the user's real-time perception needs and further enhancing the display effect and user experience of highly reliable human-computer interaction with robots. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a method for ensuring display uniformity in high-reliability human-computer interaction for robots, provided in an embodiment of the present invention. Figure 2 This is a functional block diagram of a display uniformity assurance system for high-reliability human-computer interaction in robots, provided in an embodiment of the present invention. The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] This application provides a method for ensuring display uniformity in high-confidence human-computer interaction for robots. The execution entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method for ensuring display uniformity in high-confidence human-computer interaction for robots can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cluster of cloud servers. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0020] Reference Figure 1 The diagram shown is a flowchart illustrating a method for ensuring display uniformity in high-reliability human-computer interaction for robots, according to an embodiment of the present invention. In this embodiment, the method for ensuring display uniformity in high-reliability human-computer interaction for robots includes: S1. By fusing real-time output image data and ambient light spectrum data of the display area in the robot display device, multi-dimensional display status information of the display area is obtained, and by combining the multi-dimensional display status information, the independent supplementation requirement value of the pixels in the display area is evaluated. In this embodiment of the invention, the real-time output image data and ambient light spectrum data of the display area in the fusion robot display device are used to obtain multi-dimensional display state information of the display area. Combined with this multi-dimensional display state information, the independent supplementation requirement value of pixels within the display area is evaluated, including: Collect real-time output image data and ambient light spectrum data from the display area of ​​the robot's display device; Photometric measurement processing is performed on the real-time output image data to obtain the brightness distribution characteristics of the real-time output image data; The ambient light characteristics of the ambient light spectrum data are obtained by analyzing the light intensity and color temperature parameters of the ambient light spectrum data. In the spatial and temporal dimensions, the brightness distribution features and the ambient light features are fused to obtain multidimensional display state information of the display area; Based on the multidimensional display status information, the brightness value of the pixel in the display area is compared with the preset interval reference brightness to obtain the brightness deviation data of the pixel. Based on the brightness deviation, the required brightness compensation value for the pixel is derived, and the brightness compensation value is used as the independent supplementary requirement value for the pixel.

[0021] The system collects real-time output image data and ambient light spectrum data from the display area of ​​the robot's display device. It captures frame-by-frame image information of the current display screen through the image acquisition component built into the display device, and at the same time uses the spectral sensor on the device to collect complete spectral data of the surrounding ambient light, ensuring that the two types of data are completely synchronized in terms of acquisition time and accurately correspond to the device output and environmental status at the same display moment.

[0022] Photometric measurement processing is performed on the real-time output image data. The luminous flux of each point in the image is detected pixel by pixel. The luminous flux data is converted into standard brightness units. Then, the brightness values ​​of different positions in the entire display area are statistically analyzed to clarify the distribution of brightness, concentrated areas and changing trends, and finally form the brightness distribution characteristics of the real-time output image data.

[0023] The light intensity and color temperature parameters of ambient light spectrum data are analyzed. The collected ambient light spectrum data is segmented by wavelength, and the light energy corresponding to each wavelength is accumulated to obtain the total light energy, thereby determining the light intensity parameter of ambient light. By calculating the energy ratio of red, green and blue light in the spectrum and combining the color mixing law, the color temperature parameter of ambient light is determined. The light intensity parameter and color temperature parameter are integrated to obtain the ambient light characteristics of ambient light spectrum data.

[0024] In both spatial and temporal dimensions, feature fusion is performed on brightness distribution characteristics and ambient light characteristics. Spatially, the display area is divided into multiple small spatial units, each corresponding to a set of brightness distribution characteristic data and ambient light characteristic data, establishing a correspondence between spatial location and the two types of features. Temporally, brightness distribution characteristics and ambient light characteristics at different times are recorded in the acquisition sequence, and their changes are tracked. The spatially mapped feature data is integrated with the temporal change data to obtain multi-dimensional display status information of the display area, including spatial location, brightness, ambient light intensity, ambient color temperature, and temporal change information.

[0025] Based on multi-dimensional display status information, the independent supplementary value of pixels in the display area is evaluated. The actual brightness value of each pixel is extracted from the multi-dimensional display status information. The preset interval reference brightness is a standard brightness range determined in advance based on the best visual experience of human-computer interaction. The actual brightness value of each pixel is compared with the benchmark value of the standard range, and the difference between the actual brightness value of each pixel and the benchmark value is calculated. The differences of all pixels are summarized to form the brightness deviation data of pixels in the display area.

[0026] The required brightness compensation value for each pixel is derived from the brightness deviation data. The brightness deviation data for each pixel is analyzed. If the actual brightness value of the pixel is lower than the reference value, the brightness compensation value is the positive value of the deviation. If the actual brightness value of the pixel is higher than the reference value, the brightness compensation value is the negative value of the deviation. The magnitude of the compensation value is just enough to adjust the pixel brightness to the reference value of the preset range of reference brightness. This derived brightness compensation value is directly used as the independent supplementary requirement value for the pixel.

[0027] The beneficial effects are that by synchronously collecting, analyzing and merging multi-source data, the multi-dimensional display status information of the display area can be fully obtained, and the independent supplementary requirement value of each pixel can be accurately calculated. This provides accurate and comprehensive data support for subsequent display uniformity adjustment, ensuring that the compensation adjustment can accurately match the actual needs of each pixel, and improving the accuracy of display uniformity assurance from the basic level.

[0028] S2. Based on the spatial distribution characteristics of the independent supplementary demand values, perform brightness transmission association on the pixels to obtain the visual association constraints of the pixels; In this embodiment of the invention, the step of performing luminance conduction association on the pixels based on the spatial distribution characteristics of the independent supplementary demand values ​​to obtain the visual association constraints of the pixels includes: Based on the spatial distribution characteristics of the independent supplementary demand values, the boundary information of the high-demand area and the low-demand area in the display area is obtained; Based on the boundary information, spatial continuity analysis is performed on adjacent pixels to obtain the brightness transmission relationship of the adjacent pixels; Based on the brightness transmission relationship, a visual association graph of the pixels is constructed, with the pixels as nodes, the adjacency relationships of the pixels as edges, and the brightness transmission intensity of the adjacent pixels as edge weights. Semantic extraction is performed on the visual association map to obtain the visual association constraints of the pixels.

[0029] To obtain the boundary information between high-demand and low-demand regions in the display area based on the spatial distribution characteristics of independent supplementary demand values, the following steps are required: First, a comprehensive analysis of the spatial distribution characteristics of independent supplementary demand values ​​is needed. This feature includes the independent supplementary demand value of each pixel within the display area and its corresponding spatial coordinates. By presenting this data in the form of a two-dimensional grid, the differences in demand values ​​of pixels at different locations are visually displayed. Next, a unified demand value classification standard is established. This standard is determined based on the maximum brightness adjustment capability of the display device and the user's perception threshold for brightness uniformity in human-computer interaction scenarios. Pixel sets with independent supplementary demand values ​​higher than the upper limit of this standard are designated as high-demand regions, and pixel sets with independent supplementary demand values ​​lower than the lower limit are designated as low-demand regions. Then, the entire two-dimensional grid of the display area is scanned row by row and column by column, comparing the regional attributes of adjacent pixels. When a pixel belongs to a high-demand region and its adjacent pixels belong to a low-demand region, the spatial coordinates of that pixel are recorded. All such coordinates are connected in spatial order to form continuous boundary lines. Simultaneously, the positional information of each pixel on the boundary lines and the difference in demand values ​​between adjacent pixels are marked. Finally, the boundary information between high-demand and low-demand regions in the display area is completely obtained.

[0030] Based on boundary information, spatial continuity analysis is performed on adjacent pixels to obtain the brightness transmission relationship between them. First, using the boundary information as the core, the boundary region and a transition region of 3-5 pixels wide on each side of the boundary are delineated. Simultaneously, all other areas within the display region except the transition region are covered, ensuring that all adjacent pixels are included in the analysis range. Adjacent pixels are defined as pixels directly adjacent in spatial coordinates, including horizontally adjacent (left-right) and vertically adjacent (up-down). For each pair of adjacent pixels, their independent supplementary requirement values ​​are extracted, the difference between the two requirement values ​​is calculated, and then it is observed whether there is a display discontinuity in the spatial arrangement of this pair of pixels, i.e., whether pixel positions are missing or display is invalid due to device hardware problems or abnormal display data. If the difference in required values ​​between adjacent pixels is small and the spatial arrangement is complete without any breaks, it is determined that the brightness between this pair of pixels can be smoothly transferred, forming a positive brightness transmission relationship. If the difference in required values ​​is large or there are breaks in the spatial arrangement, it is determined that the brightness transmission between this pair of pixels is hindered, forming a restricted brightness transmission relationship. At the same time, the specific reasons for hindering the transmission are recorded, such as excessive difference in required values ​​or spatial discontinuity. By analyzing each pair of adjacent pixels one by one, the complete brightness transmission relationship between adjacent pixels is obtained.

[0031] A visual association map of pixels is constructed based on luminance conduction relationships. First, each pixel within the display area is treated as an independent node in the visual association map, and each node is assigned a unique identifier containing its spatial coordinates and independent supplementary requirement values, facilitating subsequent tracing of the actual pixel attributes corresponding to the node. Next, edges in the map are determined based on pixel adjacency relationships. For any two adjacent pixel nodes, if a luminance conduction relationship exists between them, an edge is established between these two nodes; the existence of this edge indicates a luminance association between the two pixels. Then, the luminance conduction intensity of adjacent pixels is extracted from the luminance conduction relationships. The conduction intensity is determined based on the difference in requirement values ​​and spatial continuity; the smaller the difference in requirement values ​​and the better the spatial continuity, the higher the conduction intensity. This intensity value is used as the edge weight of the corresponding edge and labeled on the edge between nodes. Finally, according to the correspondence between nodes, edges, and edge weights, a hierarchical structure is used to integrate all pixel nodes, adjacent edges, and edge weights, ensuring that each node clearly displays all its adjacent nodes, corresponding edges, and edge weights, thus completely constructing the pixel visual association map.

[0032] The visual association map is semantically extracted to obtain the visual association constraints of pixels. First, the visual association map is divided into regions for analysis, extracting node association data from high-demand regions, low-demand regions, and boundary regions. The average and maximum values ​​of edge weights within each region are calculated to determine the strength benchmark of pixel brightness association in each region. Next, key association paths in the map are identified, i.e., node chains formed by continuous connections of edges with high edge weights. These paths represent core channels for smooth brightness transmission within the display area. Pixels on these paths must maintain stable brightness associations to avoid disrupting transmission continuity during adjustment. Then, weak association points in the map, i.e., node pairs with low edge weights, are analyzed. These node pairs are areas prone to display uniformity problems. It is determined that these node pairs should be prioritized for association calibration during brightness adjustment to ensure their transmission strength meets the benchmark requirements of their respective regions. Finally, the above analysis results are transformed into specific rules, including the strength range of pixel brightness associations in different regions, protection requirements for key association paths, and calibration standards for weak association points. These rules are integrated and organized to form the visual association constraints of pixels that can directly guide subsequent brightness adjustments.

[0033] The beneficial effects are that by deriving from the spatial distribution characteristics of independent supplementary demand values ​​to visual correlation constraints in a step-by-step manner, the key correlation paths and weak points of pixel brightness transmission within the display area are accurately identified. The resulting visual correlation constraints have clear regional adaptability and operational standards, which can provide specific execution basis for subsequent equalization adjustment, effectively avoid the problem of local brightness disconnection or over-correlation during the adjustment process, ensure the continuity of overall brightness transmission in the display area, and lay a solid foundation for improving display uniformity.

[0034] S3. Using the visual association constraint as a limiting condition, the independent supplementary demand value is adjusted to obtain the initial compensation parameters of the display area. In this embodiment of the invention, the step of equalizing and adjusting the independent supplementary demand value based on the visual association constraint to obtain the initial compensation parameters of the display area includes: The visual association constraint is mapped to the compensation transmission rule of the adjacent pixels; According to the compensation conduction rule, the pixel is used as a node, the adjacency relationship of the pixel is used as a conduction link, and the conduction strength in the compensation conduction rule is used as the conduction link weight to construct the compensation conduction network of the display area. The independent supplementary demand value is used as the transmission initiation amount and transmitted along the compensation transmission network to the surrounding area of ​​adjacent pixels, while receiving compensation feedback from the surrounding area. Based on the compensation feedback, the real-time compensation amount of the pixel is adjusted, and the difference stability evaluation of the real-time compensation amount is performed to obtain the stability index of the real-time compensation amount. When the stability index reaches the preset stability condition, the real-time compensation amount is used as the initial compensation parameter for the display area.

[0035] The step of adjusting the real-time compensation amount of the pixel based on the compensation feedback, and performing a difference stability evaluation on the real-time compensation amount to obtain a stability index of the real-time compensation amount includes: The difference values ​​of the real-time compensation amounts of the adjacent pixels are statistically analyzed to obtain the compensation difference set of the adjacent pixels; The distribution pattern of the compensation difference set is analyzed to obtain the global difference level of the compensation difference set; The global difference levels are rearranged in chronological order to obtain a time sequence of the global difference levels. The time series is evaluated for trend development to obtain a stable index of the real-time compensation amount.

[0036] By mapping visual association constraints to compensation transmission rules for adjacent pixels, we first clarify the core elements of visual association constraints, such as the priority of brightness transmission between adjacent pixels, the limitation of transmission intensity, and the requirement for association stability. The transmission priority is transformed into a rule for the order of compensation value transmission: adjacent pixel pairs with higher priority receive compensation value transmission first, while those with lower priority receive it later. The transmission intensity limitation is transformed into a rule for the proportion of compensation value transmission: for adjacent pixel pairs with high transmission intensity, the proportion of compensation value transmitted from one pixel to another must not be less than a specific proportion, while for those with low transmission intensity, it is transmitted at a lower proportion. The association stability requirement is transformed into a rule for limiting the difference in compensation value: for adjacent pixel pairs with high association stability, the final difference in their compensation values ​​must be controlled within a very small fixed range and cannot exceed this range. Through this transformation, the abstract visual association constraints are converted into concrete and executable compensation transmission rules for adjacent pixels.

[0037] A compensation conduction network for the display area is constructed based on compensation conduction rules. Each pixel within the display area is treated as an independent node in the network. Each node is labeled with its corresponding spatial coordinates and initial independent compensation requirement value to ensure that the node accurately corresponds to the actual pixel. Adjacency relationships are determined based on the spatial position of the pixels. For pixels that are horizontally adjacent (left-right) or vertically adjacent (up-down), a conduction link is established between the corresponding two nodes, specifying the link's connection direction and the corresponding adjacent pixel pair. The conduction strength of each adjacent pixel pair is extracted from the compensation conduction rules, and the conduction strength value is directly used as the weight of the corresponding conduction link. The higher the conduction strength, the larger the link weight value; the lower the conduction strength, the smaller the weight value. Finally, all nodes, conduction links, and corresponding weights are integrated according to the spatial distribution relationship of pixels to form a compensation conduction network covering the entire display area, enabling interaction of compensation amounts between nodes.

[0038] The independent supplementary demand value is used as the initial transmission amount and transmitted along the compensation transmission network to the surrounding area of ​​adjacent pixels. Simultaneously, compensation feedback from the surrounding area is received. Each pixel's own independent supplementary demand value is set as its initial transmission amount in the compensation transmission network. Following the transmission order and proportion in the compensation transmission rules, starting from each pixel node, the compensation value corresponding to the initial amount is transmitted to adjacent pixel nodes through the transmission link. After receiving the transmitted compensation value, the adjacent pixel nodes calculate their temporary compensation amount based on their own independent supplementary demand value. This temporary compensation amount is then used as compensation feedback information and transmitted back to the sending pixel node along the original transmission link. After receiving compensation feedback from all surrounding adjacent nodes, the sending node records the temporary compensation amount data in each feedback, forming the compensation feedback set for that node.

[0039] The difference values ​​of the real-time compensation amounts of adjacent pixels are statistically analyzed to obtain a compensation difference set for adjacent pixels. After each compensation transmission and feedback cycle, the real-time compensation amounts of all adjacent pixel nodes in the compensation transmission network are obtained. The real-time compensation amount is the current compensation value calculated by each node based on its initial amount and the received feedback amount. For each pair of adjacent nodes, the real-time compensation amount of one node is subtracted from the real-time compensation amount of the other node, and the absolute value of the result is taken as the compensation difference value of this pair of adjacent nodes. The compensation difference values ​​of all adjacent node pairs are collected one by one and arranged in order from left to right and from top to bottom according to the spatial position of the adjacent node pairs to ensure that each difference value corresponds to a specific adjacent pixel pair, thus forming the compensation difference set for adjacent pixels.

[0040] To determine the global level of difference in the compensation difference set, a distribution pattern analysis is performed. First, basic statistical calculations are conducted on all compensation difference values ​​in the set to identify the maximum and minimum values, thus defining the overall fluctuation range of the differences. Then, the arithmetic mean of all difference values ​​is calculated to reflect the average level of the differences. Simultaneously, the variance of the difference values ​​is calculated to determine the degree to which the differences deviate from the mean. Based on these statistical results, the distribution characteristics of the compensation difference values ​​are analyzed to determine whether most differences are concentrated around the mean or dispersed over a wide range, whether the overall distribution is biased towards smaller differences or contains more larger differences. This overall distribution analysis yields the global level of difference in the compensation difference set.

[0041] The global difference levels are reorganized chronologically to obtain a time-series sequence of global difference levels. Each timestamp corresponds to a specific set of calculated global difference level data, representing a compensation transmission and feedback cycle. During the reorganization process, the global difference level data corresponding to each timestamp is arranged strictly in ascending order, ensuring that data from previous time points comes first and data from later time points comes last, without any reversal of the chronological order. Simultaneously, the corresponding timestamp information is annotated next to each data point to facilitate subsequent tracing of the compensation cycle corresponding to that data, ultimately forming a time-series sequence of global difference levels arranged chronologically.

[0042] A stability index for real-time compensation is obtained by evaluating the trend development of the time series. The global difference level data at each time point in the time series is analyzed one by one, and the change in the global difference level between adjacent time points is calculated to observe the trend of change. If the change at multiple consecutive time points gradually decreases, and the global difference level at subsequent time points remains within a preset minimum fixed range, it indicates that the difference between real-time compensation quantities is continuously narrowing and the overall trend is stabilizing. If the change fluctuates significantly, or the global difference level continuously deviates from the preset range, it indicates that the real-time compensation quantity is still in an unstable state. Based on this trend judgment, a stability index is assigned to the real-time compensation quantity. The more obvious the stable trend and the longer it is maintained, the higher the stability index value, and vice versa. This is how the stability index for real-time compensation quantity is obtained.

[0043] When the stability index reaches the preset stability condition, the real-time compensation amount is used as the initial compensation parameter for the display area. The preset stability condition is that the stability index reaches a preset numerical threshold, and this state is maintained continuously for at least three compensation transmission and feedback cycles to ensure that the stable state is not a brief, accidental occurrence. When the stability index of the real-time compensation amount is detected to meet this preset condition, the transmission and adjustment process of the compensation amount is immediately stopped. At this time, the real-time compensation amount of each pixel node no longer changes significantly and is in a stable state. The real-time compensation amounts of all pixel nodes at this time are then organized according to the spatial position of the pixels in the display area to form a set containing the compensation data of all pixels in the display area. This set is the initial compensation parameter for the display area.

[0044] The beneficial effects are that by transforming visual association constraints into specific and executable compensation transmission rules, a compensation transmission network is constructed to achieve orderly transmission and feedback of compensation amounts. By combining time series analysis to analyze the stability of real-time compensation amounts, it is ensured that the final initial compensation parameters not only comply with the specifications of visual association constraints, but also meet the independent supplementation needs of each pixel in a balanced manner. This effectively avoids the problems of local overcompensation or undercompensation, and provides stable, accurate basic parameters that meet the requirements of overall visual continuity for subsequent display uniformity optimization, thereby improving the reliability and accuracy of display uniformity assurance.

[0045] S4. Using the user's current gaze point as the center, smoothly decrease the weights outwards to generate the visual saliency weight distribution of the current gaze point, and combine the visual saliency weight distribution with the initial compensation parameters to obtain the perceptual optimization parameters of the display area. In this embodiment of the invention, the step of smoothly decreasing the weights outwards from the user's current gaze point to generate a visual saliency weight distribution for the current gaze point includes: Centered on the user's current gaze point, the display area is divided into a central gaze area, an intermediate transition area, and an outer perception area; Determine the adjacent boundary regions of the central gaze area, the intermediate transition area, and the peripheral sensing area; Based on the distance from the current gaze point to the adjacent boundary region, distance-weighted interpolation is performed on the regional baseline weights of the central gaze region, the intermediate transition region, and the peripheral perception region to obtain the boundary transition weights of the adjacent boundary regions. By integrating the regional baseline weights and the boundary transition weights, the visual saliency weight distribution of the current gaze point is obtained.

[0046] The combination of the visual saliency weight distribution and the initial compensation parameters yields the perceptual optimization parameters for the display area, including: The pixel weight values ​​in the visual saliency weight distribution are weighted and fused with the corresponding pixel compensation values ​​in the initial compensation parameters to obtain the weighted compensation parameters for the pixel. The calculation formula for the weighted fusion is as follows: ; In the formula, Indicates the first Weighted compensation parameters for each pixel. This represents the first element in the visual saliency weight distribution. Visual saliency weights for each pixel. This represents the first element in the visual saliency weight distribution. Visual saliency weights for each pixel. Indicates the first Initial compensation parameters for each pixel. This represents the preset weight adjustment factor. This represents the total number of pixels within the display area. This represents the mean of the visual saliency weight distribution. The standard deviation of the visual saliency weight distribution is represented by the following: This represents the preset normalization adjustment coefficient. This represents the square root operation. This represents the summation operation; Based on the spatial gradient characteristics of the visual saliency weight distribution, the weighted compensation parameters are spatially consistent to obtain the gradient coordination parameters of the weighted compensation parameters. Based on the regional characteristics of the current gaze point, the gradient coordination parameters are fused according to regional characteristics to obtain the perceptual optimization parameters of the display area.

[0047] Centered on the user's current gaze point, the display area is divided into a central gaze zone, an intermediate transition zone, and a peripheral perception zone. First, the effective perception range of the macula of the human eye is determined. Using the radius of this range as a standard, a circular area is drawn with the user's current gaze point as the center as the central gaze zone, which is the area where the user's visual perception is clearest. Outside the central gaze zone, an annular area is drawn as the intermediate transition zone, with the boundary where visual clarity changes from clear to a significant decrease as the inner edge and the boundary where visual clarity drops to a lower level as the outer edge. The area outside the outer edge of the intermediate transition zone, covering the remaining part of the display area, is defined as the peripheral perception zone, where the user's visual clarity is lowest.

[0048] The adjacent boundary regions of the central fixation area, intermediate transition area, and peripheral perception area are determined. The adjacent boundary region between the central fixation area and the intermediate transition area is a ring-shaped area surrounding the central fixation area. The inner edge of this ring-shaped area is the outer boundary line of the central fixation area, and the outer edge is the inner boundary line of the intermediate transition area. The adjacent boundary region between the intermediate transition area and the peripheral perception area is a ring-shaped area surrounding the intermediate transition area. The inner edge of this ring-shaped area is the outer boundary line of the intermediate transition area, and the outer edge is the inner boundary line of the peripheral perception area. By determining the boundary line positions of these two ring-shaped areas, the adjacent boundary regions of the three areas are obtained.

[0049] Based on the distance from the current gaze point to the adjacent boundary region, distance-weighted interpolation is performed on the regional baseline weights of the central gaze area, intermediate transition area, and peripheral perception area to obtain the boundary transition weights of the adjacent boundary regions. First, fixed regional baseline weights are set for the three regions: the regional baseline weight of the central gaze area is set to the highest value, the regional baseline weight of the intermediate transition area is set to a value lower than that of the central gaze area, and the regional baseline weight of the peripheral perception area is set to the lowest value. Then, the straight-line distance from the current gaze point to each point on the adjacent boundary region is measured. For the boundary region between the central gaze area and the intermediate transition area, the maximum distance from the current gaze point to each point within the boundary region is first determined. The distance from a point to the gaze point is subtracted from the maximum distance, and the result is divided by the maximum distance to obtain the distance coefficient of that point. The distance coefficient is multiplied by the regional baseline weight of the central gaze area, 1 is added, the distance coefficient is subtracted, and then multiplied by the regional baseline weight of the intermediate transition area to obtain the transition weight of that point. The transition weights of each point in the boundary region between the intermediate transition area and the peripheral perception area are calculated in the same way. The transition weights of all boundary points are organized according to their spatial positions to obtain the boundary transition weights of the adjacent boundary regions.

[0050] By integrating the regional baseline weights and boundary transition weights, the visual saliency weight distribution of the current gaze point is obtained. The weights of all pixels in the central gaze area are uniformly assigned to the regional baseline weights of the central gaze area, the weights of all pixels in the peripheral receptive area are uniformly assigned to the regional baseline weights of the peripheral receptive area, and the weights of all pixels in the intermediate transition area, except for those in adjacent boundary areas, are uniformly assigned to the regional baseline weights of the intermediate transition area. Then, the boundary transition weights of adjacent boundary areas are distributed to the pixels in the boundary areas according to their spatial positions, ensuring that the changes in pixel weights from the central gaze area to the intermediate transition area and from the intermediate transition area to the peripheral receptive area are continuous and without abrupt changes. Finally, according to the spatial arrangement order of the pixels in the display area, the weights of all pixels are organized into complete distribution data to obtain the visual saliency weight distribution of the current gaze point.

[0051] The pixel weight values ​​in the visual saliency weight distribution are weighted and fused with the corresponding pixel compensation values ​​in the initial compensation parameters to obtain the pixel weighted compensation parameters. The pixel weight values ​​in the visual saliency weight distribution come from the previously generated visual saliency weight distribution of the current gaze point, and the corresponding pixel compensation values ​​in the initial compensation parameters come from the initial compensation parameters of the display area. First, the weight value of a certain pixel is taken, and a specific power operation is performed on this weight value. This power is a fixed value determined through multiple experiments during the system design phase, referring to the user's visual perception patterns in human-computer interaction scenarios and combining the brightness adjustment range of the display device. The calculated weight result is multiplied by the initial compensation value of the pixel to obtain the first part of the product result. Then, the weight values ​​of all pixels in the visual saliency weight distribution are calculated, and each weight... All weighted values ​​undergo the same specific power operation, and the results are summed to obtain a total. The sum is divided by the total number of pixels in the display area to obtain the average value. The square root of this average value is taken to obtain a normalization term. The product of the first part is divided by the normalization term to obtain the first part of the adjustment result. Next, the difference between the weight value of the pixel and the mean of the visual saliency weight distribution is calculated. The difference is divided by the standard deviation of the visual saliency weight distribution to obtain the normalized difference. The normalized difference is multiplied by the preset normalization adjustment coefficient, and then multiplied by the initial compensation value of the pixel to obtain the second part of the adjustment result. Finally, the first part of the adjustment result and the second part of the adjustment result are added together to obtain the weighted compensation parameter of the pixel. The weighted compensation parameters of all pixels are calculated in the same way.

[0052] Based on the spatial gradient characteristics of the visual saliency weight distribution, spatial consistency coordination is performed on the weighted compensation parameters to obtain the gradient coordination parameters. First, the spatial gradient characteristics of the visual saliency weight distribution are analyzed. This characteristic refers to the trend and rate of change of weight values ​​within the display area from the central gaze area to the peripheral perception area. For example, the rate of change of weight from the central gaze area to the intermediate transition area is slower, while the rate of change from the intermediate transition area to the peripheral perception area is faster. Based on this characteristic, an allowable range of difference in the weighted compensation parameters of adjacent pixels is set for different regions. Regions with slower change rates allow smaller differences, while regions with faster change rates allow slightly larger differences. Then, the difference between the weighted compensation parameters of each pixel and those of its adjacent pixels is checked one by one. If the difference exceeds the allowable range for the corresponding region, the weighted compensation parameters of that pixel are adjusted appropriately to reduce the difference to within the allowable range. During adjustment, it is ensured that the direction of parameter change is consistent with the spatial gradient characteristics of the weight distribution. After checking and adjusting all pixels, the gradient coordination parameters of the weighted compensation parameters for each pixel are obtained.

[0053] Based on the regional characteristics of the current gaze point, the gradient coordination parameters are fused according to regional characteristics to obtain the perceptual optimization parameters of the display area. First, the characteristics of each region where the current gaze point is located are clarified. The regional characteristics of the central gaze area are high user visual acuity requirements, and the brightness uniformity error needs to be controlled within a very small range. The regional characteristics of the intermediate transition area are medium user visual acuity, and the parameter transition continuity with the central gaze area and the peripheral perception area needs to be taken into account to avoid obvious brightness discontinuities. The regional characteristics of the peripheral perception area are low user visual acuity, and the allowable range of brightness uniformity error can be appropriately widened to reduce unnecessary adjustment costs. According to these characteristics, the gradient coordination parameters of the corresponding regions are adjusted. The gradient coordination parameters in the central gaze area are further optimized to control the brightness uniformity error within a preset minimum range. The gradient coordination parameters in the intermediate transition area are adjusted to connect naturally with the parameters of the central gaze area and the peripheral perception area, and the parameter changes of adjacent areas are made continuous through fine-tuning. The gradient coordination parameters in the peripheral perception area are appropriately simplified under the premise of meeting the basic uniformity requirements to avoid overcompensation. Finally, according to the spatial position of pixels in the display area, the adjusted gradient coordination parameters of each region are integrated into a complete parameter set to ensure smooth parameter transition between different regions, thus obtaining the perceptual optimization parameters of the display area.

[0054] The beneficial effects are as follows: By generating the visual saliency weight distribution that conforms to the visual law of the human eye step by step, and combining the initial compensation parameters for deep fusion and multi-round optimization, it not only ensures that the compensation parameters in the current fixation area of the user are preferentially adapted to the visual needs, but also avoids the occurrence of brightness断层 or overcompensation in the display area through spatial consistency coordination and regional characteristic fusion, so that the finally obtained perception optimization parameters can accurately improve the display uniformity, and at the same time highly match the real-time visual perception of the user, providing a high-quality and stable display foundation for high-confidence human-robot interaction, and effectively improving the user interaction experience and the accuracy of information transmission.

[0055] S5. Input the perception optimization parameters into the display device to obtain an enhanced display image of the display area, and evaluate the display quality of the enhanced display image according to the pupil response data of the user to obtain a quality quantification index of the enhanced display image. In the embodiment of the present invention, the evaluating the display quality of the enhanced display image according to the pupil response data of the user to obtain a quality quantification index of the enhanced display image includes: Collect the pupil diameter change data of the user; Perform dual-source data coupling on the pupil oscillation frequency and amplitude change data in the pupil diameter change data to obtain a comprehensive stress deviation degree of the pupil diameter change data; Map the comprehensive stress deviation degree to the enhanced display image to obtain a visual stress area distribution of the enhanced display image; Perform stationary point density analysis on the movement trajectory of the current fixation point to obtain a visual attention distribution of the enhanced display image; Perform spatial coincidence calibration on the visual stress area distribution and the visual attention distribution to obtain a high-response superposition area of the enhanced display image; Generate a quality quantification index of the enhanced display image according to the area ratio and average stress intensity of the high-response superposition area.

[0056] Collect the pupil diameter change data of the user. Real-time capture the dynamic images of the pupils of both eyes of the user through a high-precision eye tracking device supporting the display device. This device collects no less than 30 pupil images per second to ensure data continuity. Perform contour recognition on each frame of pupil image, extract the edge pixel points of the pupil, and obtain the pupil diameter value corresponding to each frame of image by calculating the diameter of the circle formed by the edge pixel points. At the same time, record the acquisition timestamp corresponding to each diameter value, and arrange the pupil diameter values at different timestamps in chronological order to form the pupil diameter change data of the user containing the corresponding relationship between time and pupil diameter.

[0057] The comprehensive stress deviation of pupil diameter variation data is obtained by dual-source data coupling of pupil oscillation frequency and amplitude variation data. First, the pupil oscillation frequency is extracted from the pupil diameter variation data, and the number of periodic fluctuations in pupil diameter per unit time is counted; the more times, the higher the oscillation frequency. Then, the amplitude variation data is extracted, and the difference between the maximum and minimum pupil diameter per unit time is calculated; the larger the difference, the more significant the amplitude variation. A unified benchmark reference value is set, including a benchmark oscillation frequency and a benchmark amplitude variation value. The deviation ratios of the actual oscillation frequency and the benchmark oscillation frequency, and the deviation ratios of the actual amplitude variation value and the benchmark amplitude variation value are calculated respectively. The two deviation ratios are added together with the same preset weight to obtain the comprehensive stress deviation value corresponding to each time point. The comprehensive stress deviation values ​​of all time points constitute the comprehensive stress deviation value of pupil diameter variation data.

[0058] By mapping the overall stress deviation to the augmented display image, the visual stress region distribution of the augmented display image is obtained. First, a correspondence is established between the pupil data acquisition timestamp and the frame sequence of the augmented display image, ensuring that each overall stress deviation data point accurately matches a specific frame of the augmented display image. Each frame of the augmented display image is divided into several rectangular sub-regions of uniform size, each sub-region corresponding to a spatial coordinate range. Based on the overall stress deviation corresponding to each frame, a stress level is assigned to each sub-region within that frame. Sub-regions with a deviation higher than a preset high threshold are marked as high-stress regions, those with a deviation between the high and low thresholds are marked as medium-stress regions, and those with a deviation lower than the low threshold are marked as low-stress regions. The stress region marking results of all frames are integrated according to spatial coordinates to form a visual stress region distribution covering the entire augmented display image and containing location information of regions with different stress levels.

[0059] The visual attention distribution of the enhanced display is obtained by performing dwell point density analysis on the movement trajectory of the current gaze point. An eye-tracking device records the real-time coordinates of the user's current gaze point on the enhanced display, and these coordinates are connected chronologically to form the gaze point movement trajectory. A dwell time threshold is set; when the gaze point stays within a certain coordinate range for a longer period than the threshold, that range is marked as a dwell point, and the center coordinates and dwell time of the dwell point are recorded. The enhanced display is divided into rectangular sub-regions of the same size as the visual stress area distribution. The number of dwell points within each sub-region is counted; the higher the number of dwell points, the higher the dwell point density of the sub-region. Attention levels are classified according to density: sub-regions with a density higher than a preset density threshold are marked as high-attention areas, those with a density between the threshold and the lower limit are marked as medium-attention areas, and those with a density lower than the lower limit are marked as low-attention areas. The attention levels and location information of all sub-regions are integrated to obtain the visual attention distribution of the enhanced display.

[0060] Spatially calibrating the distribution of visual stress areas and visual attention distributions yields the high-response overlay region for enhanced display. First, the spatial coordinate systems of the two distributions are unified to ensure complete spatial correspondence between their rectangular sub-regions. The stress level of each sub-region in the visual stress area distribution is compared with its attention level in the visual attention distribution. When a sub-region has both a high stress level and a high attention level, it is determined to be an overlapping region. All overlapping regions are connected spatially to form continuous or dispersed region blocks. The boundary coordinates and area size of each region block are marked, and these region blocks are collectively referred to as the high-response overlay region of the enhanced display.

[0061] The quality quantification index of the enhanced display image is generated based on the area ratio of the high-response overlay region and the average stress intensity. First, the total area of ​​the high-response overlay region is calculated, which is the sum of the areas of all overlapping regions. Then, this total area is divided by the overall area of ​​the enhanced display image to obtain the area ratio of the high-response overlay region. The comprehensive stress deviation of each sub-region within all high-response overlay regions is extracted from the visual stress region distribution. The arithmetic mean of these deviations is calculated to obtain the average stress intensity of the high-response overlay region. A weighting is set for both the area ratio and the average stress intensity, with the sum of their weights being 1. The area ratio is multiplied by its corresponding weight, and then the average stress intensity is multiplied by its corresponding weight to obtain a comprehensive value. This value is the quality quantification index of the enhanced display image.

[0062] The beneficial effect is that by combining user pupil response data with the spatial characteristics of the enhanced display, the display quality is evaluated from two dimensions: stress response and focus. The generated quality quantification index can accurately reflect the user's actual perception of the display, avoiding the one-sidedness of relying solely on device parameters for evaluation. This provides a basis for subsequent dynamic adjustment of compensation parameters that is in line with the user's real experience, further improving the reliability of display quality and user satisfaction in robot human-computer interaction.

[0063] S6. Based on the quality quantification index, dynamically adjust the parameter weights of the equalization adjustment.

[0064] In this embodiment of the invention, the step of dynamically adjusting the parameter weights of the equalization adjustment based on the quality quantification index includes: The quality quantification index is compared with the preset quality benchmark value to obtain the quality deviation parameter of the enhanced display screen. Based on the quality deviation parameter, confirm the weight adjustment amount of the visual association constraint and the independent supplementary demand value in the equalization adjustment; Based on the weight adjustment amount, the parameter weight configuration of the equalization adjustment is dynamically updated.

[0065] The quality deviation parameter of the enhanced display is obtained by comparing the quality quantification index with the preset quality benchmark value. The preset quality benchmark value needs to be determined by combining the user's optimal perception threshold for display uniformity in high-confidence human-computer interaction scenarios, the hardware performance limit of the display device, and the display quality adaptation standard under different ambient light scenarios. Specifically, a large number of user experience tests are conducted to collect user feedback data on display effects in different interaction scenarios. The quality quantification value range that more than 90% of users consider to have a good display effect is selected, and the median value of this range is taken as a fixed preset quality benchmark value. After obtaining the quality quantification index corresponding to the current enhanced display, the index is numerically calculated with the preset quality benchmark value. The difference between the quality quantification index and the preset quality benchmark value is obtained. If the difference is positive, it indicates that the current display quality is higher than the benchmark requirement, and the larger the difference, the more significant the degree of exceeding the benchmark. If the difference is negative, it indicates that the current display quality is lower than the benchmark requirement, and the larger the absolute value of the difference, the more obvious the gap between the display quality and the benchmark. If the difference is in a small range close to zero, it means that the current display quality basically meets the benchmark requirement. The calculated difference, the positive or negative attribute of the difference, and the corresponding gap level between the display quality and the benchmark are integrated and recorded to form a quality deviation parameter for the enhanced display screen that includes specific deviation data and a description of the degree of deviation.

[0066] Based on the quality deviation parameters, the weight adjustment amounts of visual correlation constraints and independent supplementation demand values ​​in the equalization adjustment are determined. First, the initial parameter weight configuration for equalization adjustment is clarified. In the initial state, the sum of the weights of visual correlation constraints and independent supplementation demand values ​​is fixed at 1. The initial weights are set as follows: visual correlation constraints account for 0.4, and independent supplementation demand values ​​account for 0.6. This initial configuration is based on the general display characteristics of the display device and was determined after verification through multiple basic adjustment experiments. Subsequently, the specific situation of the quality deviation parameters is analyzed. If the quality deviation parameter is negative and the gap level is severe, it indicates that the current display quality is far from meeting the benchmark requirements. In this case, the root cause of the problem needs to be determined through image detection. If the detection finds brightness banding in the display image, it indicates insufficient constraint on pixel brightness correlation. The weight adjustment amount of visual correlation constraints needs to be increased by 0.1 each time, while the weight adjustment amount of independent supplementation demand values ​​needs to be decreased by 0.1 each time. If the detection finds areas of compensation deficiency in the display image, it indicates insufficient satisfaction of pixel independent compensation requirements. The weight adjustment amount of independent supplementation demand values ​​needs to be increased by 0.1 each time, while the weight adjustment amount of visual correlation constraints needs to be decreased by 0.1 each time. If the quality deviation parameter is positive and the gap level is severe, it indicates that the display quality excessively exceeds the baseline requirements, potentially leading to resource waste due to over-adjustment. In this case, the weights currently holding higher weights should be reduced. For example, if the current visual association constraint weight is 0.7 and the independent supplementary requirement value is 0.3, then the visual association constraint weight should be reduced by 0.05, and the independent supplementary requirement value weight should be increased by 0.05. If the quality deviation parameter is close to zero, it indicates that the current weight configuration meets the display quality requirements, and the weight adjustment amount should be set to 0. The determined increases / decreases and adjustment directions of the visual association constraints and independent supplementary requirement values ​​should be organized into clear weight adjustment amounts.

[0067] The parameter weight configuration for equalization adjustment is dynamically updated based on weight adjustment amounts. First, the currently used equalization adjustment parameter weight configuration is retrieved from the system's parameter storage module. This configuration file clearly records the current weight values ​​of visual association constraints and independent supplementary demand values. Then, the current weight values ​​are adjusted and calculated according to the weight adjustment amounts. The current weight value of the visual association constraint is added to the corresponding weight adjustment amount to obtain the new weight value of the visual association constraint; the current weight value of the independent supplementary demand value is added to the corresponding weight adjustment amount to obtain the new weight value of the independent supplementary demand value. After the adjustment calculation is completed, it is necessary to verify whether the new weight values ​​meet the preset rules. That is, the sum of the two new weight values ​​must be 1, and each weight value must be within the range of 0 to 1. If the new weight values ​​exceed the range of 0-1 or the sum of the two is not 1, the weight adjustment amount needs to be fine-tuned. For example, if the calculated new weight value of the visual association constraint is 1.05 and the independent supplementary requirement value is -0.05, then the adjustment amount of the visual association constraint is reduced by 0.05 and the adjustment amount of the independent supplementary requirement value is increased by 0.05, so that the new weight values ​​are corrected to 1.0 and 0.0 respectively, ensuring compliance with the rule requirements. After successful verification, the original parameter weight configuration content is replaced with the new weight values, and the updated configuration file is re-stored in the parameter storage module. This updated configuration will be directly used in the next equalization adjustment process, completing the dynamic update of the equalization adjustment parameter weight configuration.

[0068] The beneficial effects are that by analyzing the difference between the quality quantification indicators and the preset quality benchmark values, the problems existing in the current display quality can be accurately located. Then, by combining the root causes of the problems, the weights of visual correlation constraints and independent supplementary demand values ​​can be adjusted in a targeted manner, realizing the dynamic adaptation of the balanced adjustment parameters. This avoids the problem of rigid adjustment caused by fixed weight configuration, ensuring that each adjustment can fit the actual situation of the current display quality, continuously optimizing the display uniformity, and making the display effect of the robot in the human-computer interaction process always match the user's perceived needs, further improving the credibility of the interaction and the user experience.

[0069] like Figure 2 The diagram shown is a functional block diagram of a display uniformity assurance system for high-reliability human-computer interaction of robots provided in an embodiment of the present invention.

[0070] The display uniformity assurance system 100 for high-reliability human-computer interaction in robots, as described in this invention, can be installed in an electronic device. Depending on the functions implemented, the display uniformity assurance system 100 may include a display state information acquisition module 101, a visual association constraint determination module 102, an initial compensation parameter generation module 103, a visual saliency weight generation module 104, a display quality evaluation module 105, and a parameter weight adjustment module 106. The module described in this invention can also be referred to as a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, stored in the memory of the electronic device.

[0071] In this embodiment, the functions of each module / unit are as follows: The display status information acquisition module 101 is used to fuse real-time output image data and ambient light spectrum data of the display area in the robot display device to obtain multi-dimensional display status information of the display area, and to evaluate the independent supplementation requirement value of the pixels in the display area in combination with the multi-dimensional display status information. The visual association constraint determination module 102 is used to perform brightness transmission association on the pixel based on the spatial distribution characteristics of the independent supplementary demand value to obtain the visual association constraint of the pixel. The initial compensation parameter generation module 103 is used to equalize and adjust the independent supplementary demand value based on the visual association constraint as a limiting condition to obtain the initial compensation parameters of the display area. The visual saliency weight generation module 104 is used to smoothly decrease the weights outward from the user's current gaze point to generate the visual saliency weight distribution of the current gaze point, and combine the visual saliency weight distribution and the initial compensation parameters to obtain the perception optimization parameters of the display area. The display quality assessment module 105 is used to input the perception optimization parameters into the display device to obtain the enhanced display image of the display area, and to evaluate the display quality of the enhanced display image based on the user's pupil response data to obtain the quality quantification index of the enhanced display image. The parameter weight adjustment module 106 is used to dynamically adjust the parameter weights of the equalization adjustment according to the quality quantification index.

[0072] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0073] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0074] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0075] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0076] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for ensuring display uniformity in high-reliability human-computer interaction for robots, characterized in that, The method includes: S1. By fusing real-time output image data and ambient light spectrum data of the display area in the robot display device, multi-dimensional display status information of the display area is obtained, and by combining the multi-dimensional display status information, the independent supplementation requirement value of the pixels in the display area is evaluated. S2. Based on the spatial distribution characteristics of the independent supplementary demand values, perform brightness transmission association on the pixels to obtain the visual association constraints of the pixels; S3. Using the visual association constraint as a limiting condition, the independent supplementary demand value is adjusted to obtain the initial compensation parameters of the display area. S4. Using the user's current gaze point as the center, smoothly decrease the weights outwards to generate the visual saliency weight distribution of the current gaze point, and combine the visual saliency weight distribution with the initial compensation parameters to obtain the perceptual optimization parameters of the display area. S5. Input the perception optimization parameters into the display device to obtain the enhanced display image of the display area, and evaluate the display quality of the enhanced display image based on the user's pupil response data to obtain the quality quantification index of the enhanced display image; S6. Based on the quality quantification index, dynamically adjust the parameter weights of the equalization adjustment.

2. The display uniformity assurance method for high-reliability human-computer interaction in robots as described in claim 1, characterized in that, The fusion robot display device uses real-time output image data and ambient light spectrum data of the display area to obtain multi-dimensional display state information of the display area. Based on this multi-dimensional display state information, it evaluates the independent supplementation requirement value of pixels within the display area, including: Collect real-time output image data and ambient light spectrum data from the display area of ​​the robot's display device; Photometric measurement processing is performed on the real-time output image data to obtain the brightness distribution characteristics of the real-time output image data; The ambient light characteristics of the ambient light spectrum data are obtained by analyzing the light intensity and color temperature parameters of the ambient light spectrum data. In the spatial and temporal dimensions, the brightness distribution features and the ambient light features are fused to obtain multidimensional display state information of the display area; Based on the multidimensional display status information, the brightness value of the pixel in the display area is compared with the preset interval reference brightness to obtain the brightness deviation data of the pixel. Based on the brightness deviation, the required brightness compensation value for the pixel is derived, and the brightness compensation value is used as the independent supplementary requirement value for the pixel.

3. The display uniformity assurance method for high-reliability human-computer interaction in robots as described in claim 1, characterized in that, The step of performing luminance conduction correlation on the pixels based on the spatial distribution characteristics of the independent supplementary demand values ​​to obtain the visual correlation constraints of the pixels includes: Based on the spatial distribution characteristics of the independent supplementary demand values, the boundary information of the high-demand area and the low-demand area in the display area is obtained; Based on the boundary information, spatial continuity analysis is performed on adjacent pixels to obtain the brightness transmission relationship of the adjacent pixels; Based on the brightness transmission relationship, a visual association graph of the pixels is constructed, with the pixels as nodes, the adjacency relationships of the pixels as edges, and the brightness transmission intensity of the adjacent pixels as edge weights. Semantic extraction is performed on the visual association map to obtain the visual association constraints of the pixels.

4. The display uniformity assurance method for high-reliability human-computer interaction in robots as described in claim 1, characterized in that, The step of equalizing and adjusting the independent supplementary demand value based on the visual association constraint to obtain the initial compensation parameters for the display area includes: The visual association constraint is mapped to the compensation transmission rule of the adjacent pixels; According to the compensation conduction rule, the pixel is used as a node, the adjacency relationship of the pixel is used as a conduction link, and the conduction strength in the compensation conduction rule is used as the conduction link weight to construct the compensation conduction network of the display area. The independent supplementary demand value is used as the transmission initiation amount and transmitted along the compensation transmission network to the surrounding area of ​​adjacent pixels, while receiving compensation feedback from the surrounding area. Based on the compensation feedback, the real-time compensation amount of the pixel is adjusted, and the difference stability evaluation of the real-time compensation amount is performed to obtain the stability index of the real-time compensation amount. When the stability index reaches the preset stability condition, the real-time compensation amount is used as the initial compensation parameter for the display area.

5. The display uniformity assurance method for high-reliability human-computer interaction in robots as described in claim 4, characterized in that, The step of adjusting the real-time compensation amount of the pixel based on the compensation feedback, and performing a difference stability evaluation on the real-time compensation amount to obtain a stability index of the real-time compensation amount, includes: The difference values ​​of the real-time compensation amounts of the adjacent pixels are statistically analyzed to obtain the compensation difference set of the adjacent pixels; The distribution pattern of the compensation difference set is analyzed to obtain the global difference level of the compensation difference set; The global difference levels are rearranged in chronological order to obtain a time sequence of the global difference levels. The time series is evaluated for trend development to obtain a stable index of the real-time compensation amount.

6. The display uniformity assurance method for high-reliability human-computer interaction in robots as described in claim 1, characterized in that, The step of generating a visual saliency weight distribution for the current gaze point by smoothly decreasing the weights outwards from the current gaze point includes: Centered on the user's current gaze point, the display area is divided into a central gaze area, an intermediate transition area, and an outer perception area; Determine the adjacent boundary regions of the central gaze area, the intermediate transition area, and the peripheral sensing area; Based on the distance from the current gaze point to the adjacent boundary region, distance-weighted interpolation is performed on the regional baseline weights of the central gaze region, the intermediate transition region, and the peripheral perception region to obtain the boundary transition weights of the adjacent boundary regions. By integrating the regional baseline weights and the boundary transition weights, the visual saliency weight distribution of the current gaze point is obtained.

7. The display uniformity assurance method for high-reliability human-computer interaction in robots as described in claim 1, characterized in that, The combination of the visual saliency weight distribution and the initial compensation parameters yields the perceptual optimization parameters for the display area, including: The pixel weight values ​​in the visual saliency weight distribution are weighted and fused with the corresponding pixel compensation values ​​in the initial compensation parameters to obtain the weighted compensation parameters for the pixel. The calculation formula for the weighted fusion is as follows: ； In the formula, Indicates the first Weighted compensation parameters for each pixel. This represents the first element in the visual saliency weight distribution. Visual saliency weights of pixels, This represents the first element in the visual saliency weight distribution. Visual saliency weights of pixels, Indicates the first Initial compensation parameters for each pixel. This represents the preset weight adjustment factor. This represents the total number of pixels within the display area. This represents the mean of the visual saliency weight distribution. The standard deviation of the visual saliency weight distribution is represented by the following: This represents the preset normalization adjustment coefficient. This represents the square root operation. This represents the summation operation; Based on the spatial gradient characteristics of the visual saliency weight distribution, the weighted compensation parameters are spatially consistent to obtain the gradient coordination parameters of the weighted compensation parameters. Based on the regional characteristics of the current gaze point, the gradient coordination parameters are fused according to regional characteristics to obtain the perceptual optimization parameters of the display area.

8. The method for ensuring display uniformity for high-reliability human-computer interaction in robots as described in claim 1, characterized in that, The step of evaluating the display quality of the enhanced display based on the user's pupil response data to obtain a quantitative quality index of the enhanced display includes: Collect the user's pupil diameter change data; By performing dual-source data coupling on the pupil oscillation frequency and amplitude changes in the pupil diameter change data, the comprehensive stress deviation of the pupil diameter change data is obtained; The overall stress deviation is mapped to the enhanced display screen to obtain the visual stress region distribution of the enhanced display screen; Dwell point density analysis is performed on the movement trajectory of the current gaze point to obtain the visual attention distribution of the enhanced display screen; Spatial overlap calibration is performed on the distribution of visual stress area and visual attention distribution to obtain the high response superposition area of ​​the enhanced display image; The quality quantification index of the enhanced display image is generated based on the area ratio of the high-response overlay region and the average stress intensity.

9. The display uniformity assurance method for high-reliability human-computer interaction in robots as described in claim 1, characterized in that, The step of dynamically adjusting the parameter weights of the equalization adjustment based on the quality quantification index includes: The quality quantification index is compared with the preset quality benchmark value to obtain the quality deviation parameter of the enhanced display screen. Based on the quality deviation parameter, confirm the weight adjustment amount of the visual association constraint and the independent supplementary demand value in the equalization adjustment; Based on the weight adjustment amount, the parameter weight configuration of the equalization adjustment is dynamically updated.

10. A display uniformity assurance system for high-confidence human-computer interaction in robots, used to implement the display uniformity assurance method for high-confidence human-computer interaction in robots as described in claim 1, the system comprising: The display status information acquisition module is used to fuse real-time output image data and ambient light spectrum data of the display area in the robot display device to obtain multi-dimensional display status information of the display area, and to evaluate the independent supplementation requirement value of the pixels in the display area in combination with the multi-dimensional display status information. The visual association constraint determination module is used to perform brightness transmission association on the pixel based on the spatial distribution characteristics of the independent supplementary demand value, so as to obtain the visual association constraint of the pixel. The initial compensation parameter generation module is used to equalize and adjust the independent supplementary demand value based on the visual association constraint to obtain the initial compensation parameters of the display area. The visual saliency weight generation module is used to smoothly decrease the weights outward from the user's current gaze point to generate the visual saliency weight distribution of the current gaze point, and combine the visual saliency weight distribution and the initial compensation parameters to obtain the perceptual optimization parameters of the display area. The display quality assessment module is used to input the perception optimization parameters into the display device to obtain the enhanced display image of the display area, and to evaluate the display quality of the enhanced display image based on the user's pupil response data to obtain the quality quantification index of the enhanced display image. The parameter weight adjustment module is used to dynamically adjust the parameter weights of the equalization adjustment based on the quality quantification index.

Images