A method and system for visual perception lighting for density maps

Abstract

The present application provides a kind of visualization perception lighting method and system for density map, the present application highlights ridge and valley in density field by calculating a normal field, and applies diffuse shadow to optimize lighting direction, to maximize the contrast of these structures, effectively reveal the detailed structure of high-density medium-density area and outliers of low-density area, while avoiding color distortion in color density map.The present application further considers minimizing the color distortion generated when combining shadow image and color density map, which can ensure the accuracy of density value estimation.

Description

Technical Field

[0001] This invention belongs to the field of visual perception lighting, specifically relating to a visual perception lighting method and system for density maps. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Scatter plots are one of the most effective techniques for visualizing two-dimensional discrete data points. However, for large and dense datasets, scatter plots can become overdrawn, where overlapping data points obscure the data distribution. To overcome this problem, several alternatives have been proposed: for example, instead of visualizing each point as a single marker, using a two-dimensional density map (also known as a density map or heatmap) to color-code the aggregated density of data points, typically with smoothing using kernel density estimation (KDE).

[0004] Density maps can effectively reveal overall patterns (e.g., trends and clusters), especially in high-density areas. However, they often overlook local patterns in lower-density areas, such as outliers. They can also sometimes hide important visual structures (hereinafter referred to as "structures"), i.e., meaningful local density variations. This is because existing displays and our visual systems have limited color resolution: assigning indistinguishable colors to low- to medium-density areas hinders several important analytical tasks, such as outlier identification. The following section further introduces related work on density map generation, enhancement, evaluation, and illumination models.

[0005] Currently, two common methods are used to generate density maps: opacity adjustment and density estimation. Assigning appropriate opacity to each data point can produce a semi-transparent visualization, where high-density areas can be revealed. Matejka et al. (J.Matejka, F. Anderson, and G. Fitzmaurice. Dynamic Opacity Optimization for Scatter Plots. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, pp.2707–2710. ACM, 2015) proposed a user-driven model that sets opacity values ​​based on data distribution and crowdsourced experiment results. Specifically, the authors proposed an opacity scaling model for overdrawn scatter plots, setting up a crowdsourced experiment that required users to manually set a set of controlled chart opacity levels to collect users' perceptual preferences, and then extending this to a real-world dataset. Micallef et al. (L. Micallef, G. Palmas, A. Oulasvirta, and T. Weinkauf. Towards Perceptual Optimization of the Visual Design of Scatterplots. IEEE Transactions on Visualization and Computer Graphics, 22(6):1588–1599, 2017) suggested optimizing opacity based on given dimensions, marker size, and other visual attribute data and analytical tasks. However, opacity blending is limited to a few layers, and it is difficult to distinguish different densities in the resulting density map. Instead, density estimation explicitly computes a smooth density field, which is typically displayed as a color-coded map. Density is usually computed via kernel density estimation (KDE), which is achieved by calculating the contribution of discrete data samples around each screen pixel.

[0006] Traditional continuous density plots are ineffective at describing local variations and outliers. To preserve outliers, Splatterplots (SPP), proposed by Mayorga and Gleicher (A. Mayorga and M. Gleicher. Splatterplots: Overcoming Overdraw in Scatter Plots. IEEE Transactions on Visualization and Computer Graphics, 19(9):1526–1538, 2013), explicitly displays sampled data points in sparse regions and uses closed, smooth contours in dense regions. SPPs are used to display clusters in multi-class scatter plots. They assign the same color to both high-density regions enclosed by contours and sparse regions containing similar discrete data points, hiding density variations and leading to ambiguity: the same color may correspond to different density values. On the other hand, Willems et al. (N. Willems, H. Van De Wetering, and JJ Van Wijk. Visualization of vessel movements. Computer Graphics Forum, 28(3):959–966, 2009) used a realistic Phong illumination model to plot height maps and combined them with two KDE density fields of different bandwidths to reveal local structures on color density maps. However, such illumination density maps (IDPs) cannot clearly show outliers, so the authors used a specific method to process trajectory data to highlight outliers, which is not suitable for general binary datasets. In addition, IDPs distort the original colors of the density map, hindering users from finding and comparing density values ​​based on color mapping.

[0007] Trautner et al. (T. Trautner, F. Bolte, S. Stoppel, and S. Bruckner. Sunspot Plots: Model-based Structure Enhancement for Dense Scatter Plots. Computer Graphics Forum, 39(3):551–563, 2020) introduced Sunspot Plots (SUP), which computes two density fields by performing KDE on data with two different types of kernels. After rendering these density fields using different color maps, the two rendered density fields are adaptively combined, utilizing shading and shape cues to preserve the difference in relative density. However, the color blending of SUP can introduce colors that lie outside a given color map, which hinders the user's understanding of data patterns.

[0008] Sarikaya and Gleicher (A. Sarikaya and M. Gleicher. Scatterplots: Tasks, Data, and Designs. IEEE Transactions on Visualization and Computer Graphics, 24(1):402–412, 2018) outlined a set of scatter plot-specific tasks and discussed the effectiveness of scatter plots, contour plots, and SPP in supporting aggregation-based tasks. More recently, Trautner et al. compared five visual designs and two tasks (density estimation and comparison) in a user study of SUP and found that SUP performed similarly to KDE-based continuous density plots.

[0009] Meanwhile, various lighting models have been developed. Here, we will discuss models limited to non-realistic ones, used to describe shape details and important structures rather than to create realistic images. For example, Cignoni et al. (P. Cignoni, R. Scopigno, and M. Tarini. A simple normal enhancement technique for interactive non-photorealistic renderings. Computers & Graphics, 29(1):125–133, 2005) suggested sharpening the normal field for diffuse shadows to increase corner contrast. Rusinkiewicz et al. (S. Rusinkiewicz, M. Burns, and D. DeCarlo. Exaggerated Shading for Depicting Shape and Detail. ACM Transactions on Graphics (TOG), 25(3):1199–1205, 2006) proposed a multi-scale shadow model that conveys overall shape and fine-scale details by calculating the normal field at different scales and changing the light source to maximize contrast. However, these existing models do not take into account the task requirements of density map visualization, namely, to reveal the structure and outliers on the density map while maintaining the original appearance of the color density map that matches the color mapping.

[0010] In summary, current techniques for processing density map lighting all have their limitations, and there is no widely adopted standard solution. Summary of the Invention

[0011] To address the aforementioned problems, this invention proposes a visualization-based perceptual lighting method and system for density maps. This invention highlights ridges and valleys in the density field by calculating a normal field and applies diffuse shadows optimized for the lighting direction to maximize the contrast of these structures. This effectively reveals the detailed structure of high-density regions and outliers in low-density regions, while avoiding color distortion in the color density map. Furthermore, this invention considers minimizing the color distortion generated when combining shadow images and color density maps, ensuring the accuracy of density value estimation.

[0012] According to some embodiments, the present invention adopts the following technical solution:

[0013] A visualization-based perceptual lighting method for density maps includes the following steps:

[0014] Obtain the density map;

[0015] Based on the density map, images at different scales are generated using Gaussian smoothing, and the structure map is determined based on the density difference.

[0016] The content in the structure diagram is rendered using diffuse shadows, and lighting is configured to provide an overall contrast greater than a set value to obtain a shadow intensity field;

[0017] The shadow intensity field is integrated into the luminance channel of the color density map to obtain the final enhanced density map.

[0018] As an alternative implementation, the specific process of determining the structure map based on images at different scales generated by Gaussian smoothing and density differences includes:

[0019] By measuring the density difference of Gaussian estimates with different kernel sizes, i.e., the Gaussian difference function, we can analyze images of different scales generated by Gaussian smoothing; and derive the surface normal from the gradient.

[0020] As an alternative implementation, the process of rendering the content in the structure diagram using diffuse shadows includes generating visual cues that allow for the perception of structural details using the Lambert shadow model.

[0021] Multiply the values ​​in the structure diagram by a specified factor η to exaggerate them significantly.

[0022] The details of the region of interest are interactively emphasized through the brushing operation and adjustment factor η.

[0023] As a further implementation, the specific process of multiplying the values ​​in the structure diagram by a specified factor η to achieve a high degree of exaggeration includes:

[0024]

[0025] Among them, FDoG This is a structural diagram. Let N be the partial derivatives along the x and y directions, respectively. ex The surface normal field is highly exaggerated, and η is a factor.

[0026] As an alternative implementation, configuring lighting to provide an overall contrast greater than a set value includes: calculating the normal vector N. DoG The mean and spatial variation are used to obtain statistical information about the structure in the structure diagram;

[0027] Filter out the normal vectors of empty regions in the scatter plot;

[0028] Principal component analysis is performed on the x and y dimensions of the filtered normal vector N′ to obtain the principal vectors {v1, v2} and the corresponding variances {λ1, λ2}.

[0029] Choose a lighting direction from the top where the y-component is negative, and set the pitch angle to a predetermined value.

[0030] As a further implementation, the selection of an illumination direction where the y-component emitted from the top is negative is defined by the following formula:

[0031]

[0032] Combine() generates a three-dimensional unit vector, N′ mean λ is the mean of the filtered normal vector N′, and v1 is the direction of the principal vector corresponding to the maximum variance λ1.

[0033] As an alternative implementation, the specific process of integrating the shadow intensity field into the luminance channel of the color density map includes:

[0034] Scaling the intensity field to fit the brightness range in the target color space;

[0035] Add the scaled intensity field to the luminance channel of the density map.

[0036] As a further step, the intensity field is scaled using a linear mapping, where:

[0037]

[0038] I min I is the global minimum value in the intensity field I. empty The intensity of the empty region in the scatter plot is [0, 0, 1]·L. auto φ is the brightness adjustment parameter, which controls the corresponding brightness change to I. min The mapping.

[0039] As a further step, the specific process of adding the scaled intensity field to the luminance channel of the density map includes:

[0040] LAB = RGBtoLAB(Colormap(F) large ))

[0041] LAB.L=clamp(LAB.L+I′,0,100)

[0042] PLABtoRGB(LAB)

[0043] RGBtoLAB() and LABtoRGB() are conversions between RGB and CIELAB color spaces. The clamp function is used to ensure that any invalid values ​​outside the luminance range [0,100] are adjusted to values ​​suitable for the target color space. P is the final enhanced density map.

[0044] A visualization-based perceptual lighting system for density maps, comprising:

[0045] The acquisition module is used to acquire density maps;

[0046] The Gaussian smoothing module is used to generate images at different scales based on the density map using Gaussian smoothing, and to determine the structure map based on the density difference.

[0047] The shadow intensity field construction module is used to render the content in the structure diagram using diffuse shadows and configure lighting to provide an overall contrast greater than a set value to obtain the shadow intensity field;

[0048] The shadow intensity field integration module is used to integrate the shadow intensity field into the luminance channel of the color density map to obtain the final enhanced density map.

[0049] A computer-readable storage medium for storing computer instructions, which, when executed by a processor, perform the steps in the above method.

[0050] An electronic device includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor, wherein the computer instructions, when executed by the processor, perform the steps in the method described above.

[0051] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0052] 1. This invention proposes a visual perception lighting model that supports density map-specific analysis tasks and applies a data-driven approach to automatically set lighting to capture prominent structures in the density map.

[0053] 2. This invention proposes a novel image synthesis technique that can reduce the interference between shadow images and the density values ​​of color encoding, thereby reducing color distortion;

[0054] 3. This invention can effectively reveal the detailed structure of high and medium density regions and outliers in low density regions, while avoiding color distortion in color density maps, and supports the visualization of large-scale datasets.

[0055] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0056] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0057] Figure 1 This is a flowchart of this embodiment;

[0058] Figure 2 This is a flowchart of running the method of this embodiment on an example dataset;

[0059] Figure 3 The normal vector N of the comparison density field large and N small and the normal vector N of the structure graph DoG The effect of revealing local features corresponds to, respectively Figure 2 In (a)-(c), the x and y components of the normal vector are visualized using a bivariate color map;

[0060] Figure 4 It compares different exaggeration factors (a, c)η = 1 and (b, d)η = 5 against N. ex The impact;

[0061] Figure 5 It is a comparison between two unsuitable lighting directions and two automatic lighting directions;

[0062] Figure 6 This compares traditional image compositing against both light and dark backgrounds with the brightness-only image compositing method in this embodiment. Detailed Implementation

[0063] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0064] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0065] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0066] Where there is no conflict, the embodiments and features described in this application may be combined with each other.

[0067] Example 1

[0068] A method for visualizing perceived lighting using density maps, such as Figure 1 As shown, the specific steps are as follows:

[0069] (1) Create a shadow intensity field to capture prominent structures in the density map;

[0070] (2) Develop an image synthesis method that integrates the shadow intensity field into the brightness channel of the color density map.

[0071] In this embodiment, step (1), creating a shadow intensity field to capture prominent structures in the density map, mainly includes the following steps: Figure 2 (a)-(d) are schematic diagrams of this process:

[0072] (1-1) First, analyze the images at different scales generated by Gaussian smoothing, such as Figure 2 (a) and Figure 2 As shown in (b), the density difference estimated by Gaussians for different kernel sizes is measured, and the structure diagram F is derived based on their differences. DoG To capture important structures, such as Figure 2 As shown in (c);

[0073] (1-2) Highly exaggerated diffuse shadows are used to render the content in the structure diagram, allowing for emphasis on local details in areas of interest;

[0074] (1-3) A data-driven approach is used to automatically set the lighting to provide maximum overall contrast. Combined with steps (1-2), the intensity field I can be obtained, such as... Figure 2 As shown in (d).

[0075] In step (1-1), constructing a structure diagram to capture important structures mainly includes the following steps:

[0076] (1-1-1) By measuring the density difference of Gaussian estimates with different kernel sizes, i.e. the Gaussian difference function (DoG), we can analyze images of different scales generated by Gaussian smoothing.

[0077] (1-1-2) Derive the surface normal from the gradient. Figure 3 They were displayed respectively Figure 2 of (a), Figure 2 (b) Figure 2 The x-component and y-component of the normal vector in (c).

[0078] The difference of Gaussians (DoG) function in step (1-1-1) is defined by the following formula:

[0079]

[0080] Among them, KDE h () is the Gaussian kernel function, h large and h small The bandwidths are divided into two scales, large and small, which together determine the scale of the extracted structure.

[0081] In step (1-1-2), the surface normal is derived from the gradient and defined by the following formula:

[0082]

[0083] Among them, F DoG This is a structural diagram. Let N be the partial derivatives along the x and y directions, respectively. DoG This represents the surface normal field of the structure diagram. It is related to the density field F. large or F small In comparison, structural diagram F DoG This can better enhance local structures because the relatively high-density and low-density structures in a local area are captured by the differences between density fields at different scales, and are not affected by absolute density, such as... Figure 3 As shown in the blue and green boxes.

[0084] In steps (1-2), the use of highly exaggerated diffuse shadows to render the structure diagram mainly includes the following steps:

[0085] (1-2-1) The Lambert shadow model is used to generate visual cues that allow for the perception of structural details. This model typically assumes an ideal surface that reflects light uniformly in all directions from the upper hemisphere of the surface, and that the reflection intensity I follows the Lambert cosine theorem.

[0086] (1-2-2) Although as long as it is for such Figure 5 The structural diagram N shown in (c) DoGApplying diffuse shading to the normals can reveal the detailed structure of high- and medium-density regions, but the visibility of low-density outliers remains a problem because the slope caused by outliers is often too small. For example, they are difficult to perceive. Figure 3 The outlier in the red box of (c). To resolve this issue, the value (F) in the structure diagram... DoG Multiply by a user-specified factor η to achieve a high degree of exaggeration. Figure 4 Showing η to N ex An example of the effect of different components. By increasing η from 1 to 5, the z-component of the unit normal vector is reduced, while the x and y components are correspondingly increased. Figure 4 (b) A green dot appears within the red box, while Figure 4 The outliers within the red box (d) also become more apparent;

[0087] (1-2-3) Provide users with an interactive system that uses brushing operations and adjustments to interactively emphasize details of the region of interest.

[0088] Specifically, in this embodiment, the Lambert shadow model in step (1-2-1) is defined by the following formula:

[0089] I=N·L=|N||L|cosθ=cosθ

[0090] Where N is the unit normal vector of the surface, L is the illumination direction vector from the surface to the light source, and θ is the angle between N and L.

[0091] In step (1-2-2), the value (F) in the structure diagram will be... DoG Multiplying by a user-specified factor η for high exaggeration is defined by the following formula:

[0092]

[0093] Among them, F DoG This is a structural diagram. Let N be the partial derivatives along the x and y directions, respectively. ex The surface normal field is highly exaggerated, and η is a factor.

[0094] In steps (1-3), a data-driven approach is used to automatically set the lighting to provide maximum overall contrast, mainly including the following steps:

[0095] (1-3-1) Calculate the normal vector N DoG The mean and spatial variation of F are obtained. DoG Statistical information on the structure;

[0096] (1-3-2) Filter out the normal vectors (normal vector [0,0,1]) of empty regions in the scatter plot because they do not contain valuable structure;

[0097] (1-3-3) Perform principal component analysis (PCA) on the x and y dimensions of the filtered normal vector N′ to obtain the principal vectors {v1, v2} and the corresponding variances {λ1, λ2};

[0098] (1-3-4) Select a lighting direction from the top where the y-component is negative, and set the pitch angle to 60°. Figure 5 (c) shows the shadows generated by our automatic lighting setting algorithm (azimuth = 41.44°), where the lighting direction maximizes the overall contrast and clearly shows the ridges and valleys within the blue and green boxes. Figure 5 (d) also shows the structure, but the light shining from below is not conducive to the user making the most accurate estimate of the structure.

[0099] In this embodiment, the automatic light direction in step (1-3-4) is defined by the following formula:

[0100]

[0101] Combine() generates a three-dimensional unit vector, N′ mean λ is the mean of the filtered normal vector N′, and v1 is the direction of the principal vector corresponding to the maximum variance λ1.

[0102] In this embodiment, step (2), developing an image synthesis method to integrate the shadow intensity field into the brightness channel of the color density map mainly includes the following steps:

[0103] (2-1) Figure 6 (a) shows the processing using traditional image synthesis methods. Figure 2 (d) and Figure 2 The result of (a) shows that while the structure and outliers in the blue and red boxes are clearly displayed, the original colors of the green boxes and the background have changed significantly. Furthermore, this compositing method is unsuitable for dark backgrounds because it always darkens the colors, as... Figure 6 As shown in (c). To solve this problem, the intensity field must first be scaled to fit the brightness range in the CIELAB color space while keeping the color of the empty areas in the scatter plot unchanged;

[0104] (2-2) Add the scaled intensity field to the brightness channel of the density map. For example... Figure 6As shown in (b) and (d), the brightness-only image synthesis method of this embodiment maintains the appearance of the original color by preserving the hue and saturation, and works under different backgrounds by flexibly changing φ: a negative value is used to darken the structure under a light background, while a positive value is used to brighten the structure under a dark background.

[0105] In step (2-1), scaling can be performed using a linear mapping, defined by the following formula:

[0106]

[0107] Where I min I is the global minimum value in the intensity field I. em p ty The intensity of the empty region in the scatter plot is [0, 0, 1]·L. auto φ is the brightness adjustment parameter, which controls the corresponding brightness change to I. min The mapping.

[0108] In step (2-2), adding the scaled intensity field to the luminance channel of the density map mainly includes the following formulas:

[0109] LAB = RGBtoLAB(Colormap(F) large ))

[0110] LAB.L=clamp(LAB.L+I′,0,100)

[0111] P = LABtoRGB(LAB)

[0112] RGBtoLAB() and LABtoRGB() are conversions between the RGB and CIELAB color spaces. The clamp function ensures that any invalid values ​​outside the luminance range [0, 100] are adjusted to values ​​suitable for the CIELAB color space. P is the final enhanced density map, as shown below. Figure 2 As shown in (e).

[0113] Example 2

[0114] A visualization-based perceptual lighting system for density maps, comprising:

[0115] The acquisition module is used to acquire density maps;

[0116] The Gaussian smoothing module is used to generate images at different scales based on the density map using Gaussian smoothing, and to determine the structure map based on the density difference.

[0117] The shadow intensity field construction module is used to render the content in the structure diagram using diffuse shadows and configure lighting to provide an overall contrast greater than a set value to obtain the shadow intensity field.

[0118] The shadow intensity field integration module is used to integrate the shadow intensity field into the luminance channel of the color density map to obtain the final enhanced density map.

[0119] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0120] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0121] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0122] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0123] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made by those skilled in the art without creative effort within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A visualization-based perceptual lighting method for density maps, characterized in that, Includes the following steps: Obtain the density map; Based on the density map, images at different scales are generated using Gaussian smoothing, and the structure map is determined based on the density difference. The specific process includes: By measuring the density difference of Gaussian estimates with different kernel sizes, i.e., the Gaussian difference function, we can analyze images of different scales generated by Gaussian smoothing; and derive the surface normal from the gradient. The content in the structure diagram is rendered using diffuse shadows, and lighting is configured to provide an overall contrast greater than a set value to obtain a shadow intensity field; The specific process of multiplying the values ​​in the structure diagram by a specified factor η to achieve a high degree of exaggeration includes: ； in, This is a structural diagram. , These are the partial derivatives along the x and y directions, respectively. This is the surface normal field after being highly exaggerated. η As a factor; The shadow intensity field is integrated into the luminance channel of the color density map to obtain the final enhanced density map. The specific process of integrating the shadow intensity field into the luminance channel of the color density map includes: scaling the intensity field to fit the luminance range in the target color space; and adding the scaled intensity field to the luminance channel of the density map, specifically including: RGBtoLAB() and LABtoRGB() are conversions between RGB and CIELAB color spaces. The clamp function ensures that any invalid values ​​outside the luminance range [0,100] are adjusted to values ​​suitable for the target color space. This is the final enhanced density map.

2. The visualization-based perceptual lighting method for density maps as described in claim 1, characterized in that, The process of rendering content in a structure diagram using diffuse shadows involves using the Lambert shadow model to generate visual cues that allow for the perception of structural details. Multiply the values ​​in the structure diagram by a specified factor η to exaggerate them significantly. The details of the region of interest are interactively emphasized through the brushing operation and adjustment factor η.

3. The visualization-based perceptual lighting method for density maps as described in claim 1, characterized in that, The specific process of configuring lighting to provide an overall contrast greater than a set value includes: calculating the normal vector. The mean and spatial variation are used to obtain statistical information about the structure in the values ​​of the structure diagram; Filter out the normal vectors of empty regions in the scatter plot; For the filtered normal vector Principal component analysis is performed on the x and y dimensions to obtain the principal vectors. and the corresponding variance ; Choose a lighting direction from the top where the y-component is negative, and set the pitch angle to a predetermined value.

4. The visualization-based perceptual lighting method for density maps as described in claim 3, characterized in that, The selection of a lighting direction from which the y-component of the emitted light is negative is defined by the following formula: Combine() generates a three-dimensional unit vector. It is the filtered normal vector The mean, It is the maximum variance The corresponding principal vector direction.

5. The visualization-based perceptual lighting method for density maps as described in claim 4, characterized in that, The intensity field is scaled using a linear mapping, where: This is the global minimum value in intensity field I. The intensity of the empty region in the scatter plot, i.e. , These are brightness adjustment parameters, used to control the corresponding brightness changes. The mapping.

6. A visualization-based perceptual lighting system for density maps, characterized in that, include: The acquisition module is used to acquire density maps; The Gaussian smoothing module is used to generate images at different scales based on the density map using Gaussian smoothing, and to determine the structure map based on the density difference. The specific process includes: By measuring the density difference of Gaussian estimates with different kernel sizes, i.e., the Gaussian difference function, we can analyze images of different scales generated by Gaussian smoothing; and derive the surface normal from the gradient. The shadow intensity field construction module is used to render the content in the structure diagram using diffuse shadows and configure lighting to provide an overall contrast greater than a set value to obtain the shadow intensity field. The specific process of multiplying the values ​​in the structure diagram by a specified factor η to achieve a high degree of exaggeration includes: ； in, This is a structural diagram. , These are the partial derivatives along the x and y directions, respectively. This is the surface normal field after being highly exaggerated. η As a factor; The shadow intensity field integration module is used to integrate the shadow intensity field into the luminance channel of the color density map to obtain the final enhanced density map. The specific process of integrating the shadow intensity field into the luminance channel of the color density map includes: scaling the intensity field to fit the luminance range in the target color space; and adding the scaled intensity field to the luminance channel of the density map, specifically including: RGBtoLAB() and LABtoRGB() are conversions between RGB and CIELAB color spaces. The clamp function ensures that any invalid values ​​outside the luminance range [0,100] are adjusted to values ​​suitable for the target color space. This is the final enhanced density map.

Images