A method for evaluating the complexity of calligraphy works based on layout aesthetics

Abstract

The invention discloses a method for evaluating the complexity of calligraphy works based on layout aesthetics, including obtaining the size, shape, distance, slope characteristics and grid blank characteristics of characters, sorting the above characteristics according to the order in which the author writes the characters to obtain sequence characteristics, obtaining the stationary distribution probability of the reduction, stability and increase state of each sequence characteristic based on the Markov chain principle, adding the stationary distribution probability of the reduction and increase state to obtain a complexity index, and evaluating the complexity of the calligraphy works based on the obtained size complexity index, shape complexity index, distance complexity index, slope complexity index and grid blank complexity index. The method can more accurately evaluate the complexity of the calligraphy works.

Description

Technical Field

[0001] The present invention belongs to the interdisciplinary field of computer technology and design, and specifically relates to a method for evaluating the complexity of calligraphy works based on layout aesthetics. Background Art

[0002] Chinese calligraphy is an ancient art of writing Chinese characters. From oracle bone inscriptions, stone drum inscriptions, bronze inscriptions (bell and tripod inscriptions) to large seal script, small seal script, official script, and cursive script, regular script, and running script that were finalized in the Eastern Han, Wei, and Jin dynasties, calligraphy has always exuded the charm of art. Chinese calligraphy is a unique performance art created by the Han people. It is praised as: poetry without words, dance without steps; painting without pictures, music without sound. Chinese characters are an important factor in Chinese calligraphy, because Chinese calligraphy was created and developed in Chinese culture, and Chinese characters are one of the basic elements of Chinese culture. Relying on Chinese characters is the main feature that distinguishes Chinese calligraphy from other types of calligraphy.

[0003] The writing of words has developed to an aesthetic stage - incorporating the creator's ideas, thinking, and spirit, and can inspire the aesthetic emotions of the aesthetic object (that is, the formation of calligraphy in the true sense). The earliest records of this were between the late Han Dynasty and the Wei and Jin Dynasties (approximately the second half of the 2nd century to the 4th century AD). However, this does not mean ignoring, downplaying, or even denying the artistic value and historical status of the previous calligraphy art form. The origin of Chinese characters and the creation of early works with initial artistic qualities all have their own particularity and timeliness. In terms of calligraphy, although the early Chinese characters - oracle bone inscriptions, and pictographic characters, the same character has different levels of complexity and strokes. However, it already has the rules of symmetry and balance, as well as some regular factors in the use of pens (knives), word formation, and composition. Moreover, in terms of the organization of lines and the changes in the start and end of strokes, it already has the meaning of ink writing and the significance of brushwork. Therefore, it can be said that the creation and existence of previous calligraphy art not only belongs to the category of calligraphy history, but is also an important example that can be used as a reference and reflection in the development and evolution of art forms in later generations.

[0004] Chinese calligraphy is one of the symbols of cultural heritage, and its aesthetic value has been widely used in visual design, such as movie posters and modern fashion. In human-computer interaction, the complexity of visual design is considered to have a great influence on user preferences, and the literature Carballal, A., Castro, L., Fernandez-Lozano, C., Rodr′1guez-Fern′andez, N., Romero, J., and Machado, P. (2019). Aesthetic composition indicator based on image complexity. In Interface Support for Creativity, Productivity, and Expression in Computer Graphics, pages 185–202. IGI Global. and Fernandez-Lozano, C., Carballal, A., Machado, P., Santos, A., and Romero, J. (2019). Visual complexity modelling based on image features fusion of multiple kernels. Peer J, 7: e7075. The above literature has disclosed the correlation between visual complexity and aesthetic preference. However, the visual complexity of calligraphy art has rarely been studied. Summary of the invention

[0005] The present invention provides a method for evaluating the complexity of calligraphy works based on layout aesthetics. The method can more accurately evaluate the complexity of calligraphy works.

[0006] A method for evaluating the complexity of calligraphy works based on layout aesthetics, comprising:

[0007] (1) Obtaining a bounding box of each character in the calligraphy work, and obtaining the size, shape, distance, and slope features of each character based on the bounding box of each character;

[0008] (2) According to the order in which the author writes the characters, the size, shape, distance and slope features of each character and each grid blank feature are time-sorted to obtain the size sequence feature, shape sequence feature, distance sequence feature, slope sequence feature and grid blank sequence feature;

[0009] (3) A size complexity index is obtained by using a complexity index obtaining method based on the size sequence characteristics. The specific steps of the complexity index obtaining method are as follows:

[0010] Each character size feature in the size sequence feature is used as a Markov chain value, and the state corresponding to each Markov chain value includes a decreasing, stable or increasing state. A state transfer matrix is ​​constructed through the state transfer probability distribution. Based on the state transfer matrix, the equation is solved using the Markov chain stationary distribution to obtain the state stationary distribution probability of the size sequence feature. The decreasing and increasing state stationary distribution probabilities are added to obtain the size complexity index; wherein the state transfer probability distribution is the probability distribution of the state at the previous moment transferring to the state at the current moment;

[0011] (4) Adopting the complexity index acquisition method in step (3), the shape complexity index, distance complexity index, slope complexity index and grid blank complexity index are obtained based on the shape sequence characteristics, distance sequence characteristics, slope sequence characteristics and grid blank sequence characteristics respectively, and the complexity of the calligraphy work is evaluated based on the size complexity index, shape complexity index, distance complexity index, slope complexity index and grid blank complexity index.

[0012] The step of constructing a bounding box of each character in the calligraphy work includes:

[0013] The height and width of the boundary box are the height and width of the character respectively; the horizontal and vertical coordinates of the boundary box are the horizontal and vertical coordinates of the upper left corner of the character.

[0014] According to the order in which the author writes the characters, the size features of each character are sorted in time to obtain the size sequence feature F 1 for:

[0015]

[0016] in, is the character size feature of the corresponding character at the tth moment, w t is the character width of the corresponding character at time t, h t is the character height of the corresponding character at the tth moment.

[0017] The shape features of each character are sorted in time according to the order in which the author writes the characters to obtain the shape sequence feature F 2 for:

[0018]

[0019] in, is the character shape feature at the tth moment, w t is the character width of the corresponding character at time t, h t is the character height of the corresponding character at the tth moment.

[0020] According to the order in which the author writes the characters, the distance feature of each character is sorted in time to obtain the distance sequence feature F 3 for:

[0021]

[0022] in, is the character distance feature at the tth moment, w t is the character width of the corresponding character at time t, h t is the character height of the corresponding character at the tth moment, x t is the horizontal coordinate of the character bounding box corresponding to the character at time t, y t is the ordinate of the character bounding box corresponding to the character at the tth moment.

[0023] According to the order in which the author writes the characters, the slope features of each character are sorted in time to obtain the distance sequence feature F 4 for:

[0024]

[0025] in, is the character slope feature at the tth moment, x t is the horizontal coordinate of the character bounding box corresponding to the character at time t, y t is the ordinate of the character bounding box corresponding to the character at the tth moment.

[0026] The blank features of each grid are sorted in time according to the order in which the author writes the characters to obtain the grid blank sequence features, where:

[0027] The specific steps of obtaining the blank feature of each grid are as follows: firstly, the calligraphy work is grid-cut to obtain multiple grids, and secondly, the ratio of the number of white pixels in each grid to the total pixels of the grid is used as the blank feature of each grid;

[0028] The grid blank sequence feature F 5 for:

[0029]

[0030] in, is the grid blank feature of the corresponding grid at the tth moment, V t is the number of white pixels in the grid at the tth moment, K t is the total number of pixels in the grid corresponding to the tth moment.

[0031] The state transfer matrix P is constructed by the state transfer probability distribution as follows:

[0032]

[0033] Among them, p mnis the transition probability value of the Markov chain value from state m to state n, m = 1, 2, 3 represent decreasing, stable and increasing states respectively; n = 1, 2, 3 represent decreasing, stable and increasing states respectively;

[0034] When the size complexity index is obtained by using the complexity index acquisition method, the Markov chain value is each character size feature in the size sequence feature;

[0035] When the shape complexity index is obtained by using the complexity index acquisition method, the Markov chain value is each character shape feature in the shape sequence feature;

[0036] When the distance complexity index is obtained by using the complexity index acquisition method, the Markov chain value is each character distance feature in the distance sequence feature;

[0037] When the slope complexity index is obtained by using the complexity index acquisition method, the Markov chain value is the slope feature of each character in the slope sequence feature;

[0038] When the complexity index obtaining method is used to obtain the grid blank complexity index, the Markov chain value is each grid blank feature in the grid blank sequence feature.

[0039] The Markov chain stationary distribution equation is used to solve the state transfer matrix to obtain the stationary distribution probabilities of the decreasing, stable and increasing states of the size sequence characteristics respectively. The Markov chain stationary distribution equation is:

[0040]

[0041] Among them, μ n is the stationary distribution probability of the size sequence feature in the nth state, and t is the index of the moment.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] The present invention first constructs five sequence features of each character, and based on the five sequence features of each character, uses Markov chain to obtain the state transition probability distribution of the five sequence features corresponding to the character at the current moment and the five sequence features corresponding to the character at the previous moment, constructs a state transfer matrix through the state transfer probability distribution, uses Markov chain stationary distribution to solve the equation based on the state transfer matrix to obtain the state stationary distribution probability of the size sequence feature, adds the stationary distribution probabilities of the reducing and increasing states to obtain the complexity index, and evaluates the degree of change of the five characteristic states of the character through the stationary distribution probabilities of the reducing and increasing states to more accurately evaluate the complexity of the calligraphy work. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1A schematic diagram of a calligraphy work complexity assessment method based on layout aesthetics provided by an embodiment of the present invention;

[0045] Figure 2 A schematic diagram showing calligraphy works of different scales (groups A, B, and C) provided in an embodiment of the present invention;

[0046] Figure 3 A schematic diagram of a calligraphy work complexity evaluation process provided by an embodiment of the present invention;

[0047] Figure 4 A schematic diagram showing some calligraphy images and their corresponding complexity scores provided by an embodiment of the present invention;

[0048] Figure 5 This is a schematic diagram of the analysis results of the visual complexity score provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0049] The present invention provides a method for evaluating the complexity of calligraphy works based on layout aesthetics, comprising:

[0050] S1: Extract layout features and detect the bounding box of each character in the calligraphy work, such as Figure 1 As shown in (a) in the figure, the specific steps are:

[0051] like Figure 1 As shown in (c) in the figure, each bounding box is determined by four parameters (x, y, w, h), where the height h and width w of the bounding box are the height and width of the character respectively; the horizontal and vertical coordinates (x, y) of the bounding box are the horizontal and vertical coordinates of the upper left corner of the character. The artist has a completely specific sequence in the creative process, namely the "timeline", such as Figure 1 As shown in (e), all bounding boxes are constructed into a vector [(x, y, w, h)] according to the time order of the characters written by the author. 1 ,(x,y,w,h) 2 ,…,(x,y,w,h) t ], where t is the index of the moment, and each moment corresponds to a character.

[0052] The size, shape, distance and slope features of each character are obtained based on the bounding box of each character. The size, shape, distance and slope features of each character are time-sorted according to the order in which the author writes the characters to obtain size sequence features, shape sequence features, distance sequence features and slope sequence features.

[0053] Among them, the size sequence feature F 1 for:

[0054]

[0055] in, is the character size feature of the corresponding character at the tth moment, w t is the character width of the corresponding character at time t, h t is the character height of the corresponding character at the tth moment.

[0056] The shape sequence feature F provided by this application 2 for:

[0057]

[0058] in, is the character shape feature at the tth moment, w t is the character width of the corresponding character at time t, h t is the character height of the corresponding character at the tth moment.

[0059] The distance sequence feature F provided by this application 3 for:

[0060]

[0061] in, is the character distance feature at the tth moment, w t is the character width of the corresponding character at time t, h t is the character height of the corresponding character at the tth moment, x t is the horizontal coordinate of the character bounding box corresponding to the character at time t, y t is the ordinate of the character bounding box corresponding to the character at the tth moment.

[0062] The distance sequence feature F provided by this application 4 for:

[0063]

[0064] in, is the character slope feature at the tth moment, x t is the horizontal coordinate of the character bounding box corresponding to the character at time t, y t is the ordinate of the character bounding box corresponding to the character at the tth moment.

[0065] This application uses grid sampling technology to divide the calligraphy works into 36 grids, such as Figure 1 As shown in (b), the grid parameters are expressed as vectors [grid 1 ,grid 2 ,…,grid t-1 ,grid t ], the grid corresponding to the tth moment is the 36th grid, such as Figure 1 As shown in (d) in .

[0066] The grid blank sequence feature F 5 for:

[0067]

[0068] in, is the grid blank feature of the corresponding grid at the tth moment, V t is the number of white pixels in the grid at the tth moment, K t is the total number of pixels in the grid corresponding to the tth moment.

[0069] S2: Markov chain feature modeling. Figure 1 As shown in (f) in the figure, the Markov chain is a discrete random variable S t ,t≥1 set, S t for or Satisfy p(S t+1 ∣S t ,…,S 1 )=p(S t+1 ∣S t ). The values ​​of random variables are all in the countable set S = Ω t In the above example, S is called a Markov chain, Ω represents the state space, and the value of the Markov chain in the state space corresponds to three different states, namely, decrease (T 1 ), stable (T 2 ) or increase the state (T 3 ). The probability distribution of the state of the Markov chain at time t+1 depends only on the state at time t and has nothing to do with the state before time t.

[0070] When the state corresponding to the value of the Markov chain is stable, the current value S t With the previous value S t-1 If the difference is equal to 0, it is considered that the previous state is the same as the current state; when the state corresponding to the value of the Markov chain is decreasing, the current value S t With the previous value S t-1 The difference is less than 0. When the state corresponding to the value of the Markov chain is increasing, the current value S t With the previous value S t-1 If the difference is greater than 0, it is considered that the previous moment of the current state is different from the current moment. When t = 1, it is different from S 1 The object of comparison is the average value of the corresponding characteristics of the calligraphy works.

[0071] The probability distribution of transferring the previous state to the current state is taken as the state transfer probability distribution, and the state transfer matrix P is constructed by the state transfer probability distribution:

[0072]

[0073] Among them, p mn is the transition probability value of the Markov chain value from state m to state n, m=1, 2, 3 represent decreasing, stable and increasing states respectively; n=1, 2, 3 represent decreasing, stable and increasing states respectively.

[0074] The steps for obtaining the Markov chain value provided in this application are:

[0075] When the shape complexity index is obtained by using the complexity index acquisition method, the Markov chain value is each character shape feature in the shape sequence feature;

[0076] When the distance complexity index is obtained by using the complexity index acquisition method, the Markov chain value is each character distance feature in the distance sequence feature;

[0077] When the slope complexity index is obtained by using the complexity index acquisition method, the Markov chain value is the slope feature of each character in the slope sequence feature;

[0078] When the complexity index obtaining method is used to obtain the grid blank complexity index, the Markov chain value is each grid blank feature in the grid blank sequence feature.

[0079] Based on the state transfer matrix, the Markov chain stationary distribution equation is used to solve the stationary distribution probabilities of the decreasing, stable and increasing states of the sequence characteristics. The Markov chain stationary distribution equation is:

[0080]

[0081] Among them, μ n is the stationary distribution probability of the sequence feature in the nth state, and t is the index of the moment. The complexity index is obtained by summing the stationary distribution probabilities of the decreasing and increasing states.

[0082] This application provides a method for obtaining complexity indicators for different sequence features, which is specifically described as follows:

[0083] (1) A size complexity index is obtained by using a complexity index obtaining method based on size sequence characteristics. The specific steps of the complexity index obtaining method are as follows:

[0084] Each character size feature in the size sequence feature is used as a Markov chain value, and the state corresponding to each Markov chain value includes a decreasing, stable or increasing state. A state transfer matrix is ​​constructed through the state transfer probability distribution. Based on the state transfer matrix, the equation is solved using the Markov chain stationary distribution to obtain the stationary distribution probability of the decreasing, stable and increasing states of the size sequence feature. The stationary distribution probabilities of the decreasing and increasing states are added to obtain the size complexity index; wherein the state transfer probability distribution is the probability distribution of the state at the previous moment transferring to the state at the current moment;

[0085] (2) The complexity index acquisition method in step (1) is used to obtain the shape complexity index, distance complexity index, slope complexity index and grid blank complexity index based on the shape sequence characteristics, distance sequence characteristics, slope sequence characteristics and grid blank sequence characteristics respectively, and the complexity of the calligraphy work is evaluated based on the size complexity index, shape complexity index, distance complexity index, slope complexity index and grid blank complexity index.

[0086] Example 1

[0087] A total of 85 typical calligraphy works created by 21 famous artists from the Wei Dynasty to the Qing Dynasty in China (220 AD to 1912 AD) were collected. 55 participants (20 males and 25 females) with an average age of 22.5 years (SD = 2.83) were included. Most participants had a background in art and design, and some participants already had a degree in this field.

[0088] The experimental process includes sample preparation, complexity selection and score calculation. In order to avoid the visual distraction effect of images at different scales, the collected calligraphy works are divided into dataset A (16 samples), dataset B (21 samples) and dataset C (48 samples) according to the scale of the image. Figure 2 shown.

[0089] In the sample preparation phase, three samples are randomly selected from each group as triplets. The selected triplets are presented to the participants in the selection phase. Taking dataset A as an example, Figure 3 As shown in (a) of the figure, the total number of selectable triplets is 455, but Figure 3 As shown in (b) of Figure 1, such sampling results will lead to many samples being repeatedly selected. Therefore, through an adaptive sampling method, 15 triplets are selected from data set A, 21 triplets are selected from data set B, and 48 triplets are selected from data set C. In this process, the number of repeated selections of samples is controlled so that each sample is selected 3 times on average. This process is shown in Figure 3 (c) in the figure. Participants were asked to rank the complexity of the three images as low, medium, or high. Figure 3 As shown in (d) in .

[0090] In the selection phase, calligraphy images were presented to participants as visual stimuli on a web page. Participants completed the experiment on the web page using a laptop or desktop computer. They were asked to rank the works according to their overall perception of the calligraphy layout.

[0091] In the calculation stage, three experiments were conducted on different sets and the frequency of complexity selection was calculated. The statistical histograms of some samples (samples 1, 2, 5, 11, 12, 13) in dataset A are shown in the figure below. Figure 4 To calculate the complexity score, three values ​​(x, y, z) are calculated using the Bayesian estimation method. The x value represents the probability that the sample is scored as "low complexity" in the user experiment, the y value represents the probability that it is scored as "medium complexity", and the z value represents the probability that it is scored as "high complexity".

[0092] The 7-point Likert scale is one of the most commonly used score summaries. Use 1, 4, and 7 points to represent low, medium, and high complexity scores. Therefore, the probability distribution of complexity can be converted to a numerical score as shown in the equation:

[0093]

[0094] Some comparative results of the above calculation process have been shown in Figure 4 As can be seen from the figure, it corresponds to human subjective perception. In order to obtain a more rigorous result, the following calculation process is designed:

[0095] The linear regression model is used to model the calligraphy image datasets of different shapes of the previously mentioned A, B, and C datasets. Three linear regression models are established for the experiment, namely model A, model B, and model C, corresponding to the names of their training sample sets. The model uses the complexity score of the sample as the dependent variable, and the layout feature F 1 to F 5 As the independent variable. The linear regression model fitting formula is as follows:

[0096] Y=β 0 +β 1 X 1 +…+β n X n +μ

[0097] In the formula, Y represents the dependent variable, which is the user's rating result, and X 1 …X n It represents the independent variables of each layout feature, β 0 …β n is the coefficient to be fitted, and μ is the random error. After the model is established, three indicators of the model are extracted respectively: R 2It represents the ratio of the regression sum of squares to the total sum of squares, reflecting the explanatory power of the dependent variable regression equation; F represents the F statistic; Sig. represents the p-value of the model. If it is less than 0.05, it indicates significance.

[0098] The results in Table 1, Table 2, and Table 3 show that Model A, Model B, and Model C can explain 49.9%, 65.8%, and 30.2% of the variance in the independent variables, and these models are statistically significant (Sig. < 0.05). 4 and F 5 The independent variables were statistically significant (p<0.05); F 1 、F 4 and F 5 It was statistically significant in model B (p<0.05); F 2 、F 4 and F 5 It is statistically significant (p<0.05) in Model C. This suggests that these factors can affect human perception of the visual complexity of Chinese calligraphy works.

[0099] The results of the regression coefficients in Table 1 show that in Model A, F 4 and F 5 It has a positive impact on visual complexity (regression coefficient greater than 0). This means that layouts with higher feature variance (F 4 and F 5 ) calligraphy works tend to be more complex. Similarly, the results of Model B and Model C can also be analyzed accordingly according to Tables 2 and 3. These results can prove that layout features affect visual complexity.

[0100] Table 1 shows the visual complexity analysis results of linear regression model A for data set A.

[0101]

[0102] Table 2 shows the visual complexity analysis results of linear regression model B for data set B

[0103]

[0104] Table 3 shows the visual complexity analysis results of the linear regression model C3 for the C dataset

[0105]

[0106] In order to explore the correlation between calligraphy style and visual complexity, the scores of the three calligraphy styles in dataset C were counted (regular script sample is 14, running script sample is 26, and cursive script sample is 8). The statistical results of the visual complexity score are shown in the figure below: Figure 5As shown in the figure. As can be seen from the figure, the regular script scored the lowest (mean = 3.449, SE = 0.788); the cursive script scored the highest (mean = 5.315, SE = 0.447); and the running script obtained a medium score (mean = 4.775, SE = 0.684). The standard error between the scores of the cursive script is smaller than that of the regular script and the running script. This indicates that the cursive script generally has a high degree of visual complexity in Chinese calligraphy. In summary, these results can prove that calligraphy style affects visual complexity.

[0107] The beneficial effects of the present invention are:

[0108] (1) A method for calculating the visual complexity of calligraphy based on layout aesthetic evaluation is proposed. It can extract the layout features of the input calligraphy image and map them into the visual complexity of calligraphy.

[0109] (2) The invention defines the complexity of calligraphy works and their evaluation indicators.

[0110] (3) The complexity calculation formula in the calligraphy work complexity calculation method has been proven to be significantly effective by user experiments.

Claims

1. A method for evaluating the complexity of calligraphy works based on layout aesthetics. It is characterized in that include: (1) Obtain the bounding box of each character in the calligraphy work, and obtain the size, shape, distance and slope features of each character based on the bounding box of each character; (2) The size, shape, distance, and slope features of each character, as well as each grid blank feature, are sorted in time according to the order in which the author writes the characters to obtain size sequence features, shape sequence features, distance sequence features, slope sequence features, and grid blank sequence features; (3) A size complexity index is obtained by using a complexity index obtaining method based on size sequence characteristics. The specific steps of the complexity index obtaining method are as follows: Each character size feature in the size sequence feature is used as a Markov chain value. The state corresponding to each Markov chain value includes a decreasing, stable or increasing state. A state transfer matrix is ​​constructed through the state transfer probability distribution. The state transfer probability distribution is the probability distribution of the state at the previous moment transferring to the state at the current moment. Based on the state transfer matrix, the equation is solved using the Markov chain stationary distribution to obtain the state stationary distribution probability of the size sequence feature. The decreasing and increasing state stationary distribution probabilities are added to obtain the size complexity index. (4) Adopting the complexity index acquisition method in step (3), the shape complexity index, distance complexity index, slope complexity index and grid blank complexity index are obtained based on the shape sequence features, distance sequence features, slope sequence features and grid blank sequence features respectively, and the complexity of the calligraphy work is evaluated based on the size complexity index, shape complexity index, distance complexity index, slope complexity index and grid blank complexity index.

2. According to the method for evaluating the complexity of calligraphy works based on layout aesthetics according to claim 1, It is characterized in that Constructing a bounding box for each character in a calligraphy work includes: the height and width of the bounding box are the height and width of the character respectively; and the horizontal and vertical coordinates of the bounding box are the horizontal and vertical coordinates of the upper left corner of the character.

3. The calligraphy work complexity evaluation method based on layout aesthetics according to claim 1, It is characterized in that According to the order in which the author writes the characters, the size features of each character are sorted in time to obtain the size sequence features. for: ,in, For the t The character size characteristics of the character corresponding to the moment, For the t The character width of the character corresponding to the moment, For the t The character height of the character corresponding to the moment.

4. The calligraphy work complexity evaluation method based on layout aesthetics according to claim 1, It is characterized in that The shape features of each character are sorted in time according to the order in which the author writes the characters to obtain the shape sequence features. for: ,in, For the t Character shape characteristics of the moment, For the t The character width of the character corresponding to the moment, For the t The character height of the character corresponding to the moment.

5. The calligraphy work complexity evaluation method based on layout aesthetics according to claim 2, It is characterized in that According to the order in which the author writes the characters, the distance features of each character are sorted in time to obtain the distance sequence features. for: ,in, For the t Character distance features at the moment, For the t The character width of the character corresponding to the moment, For the t The character height of the corresponding character at the moment, For the t The horizontal coordinate of the character bounding box of the character corresponding to the moment, For the t The ordinate of the character bounding box of the character corresponding to the moment.

6. The calligraphy work complexity evaluation method based on layout aesthetics according to claim 2, It is characterized in that According to the order in which the author writes the characters, the slope features of each character are sorted in time to obtain the distance sequence feature for: ,in, For the t Character slope characteristics at the moment, For the t The horizontal coordinate of the character bounding box of the character corresponding to the moment, For the t The ordinate of the character bounding box of the character corresponding to the moment.

7. The calligraphy work complexity evaluation method based on layout aesthetics according to claim 1, It is characterized in that The blank features of each grid are sorted in time according to the order in which the author writes the characters to obtain the grid blank sequence features, where: The specific steps of obtaining the blank feature of each grid are as follows: firstly, the calligraphy work is grid-cut to obtain multiple grids, and secondly, the ratio of the number of white pixels in each grid to the total pixels of the grid is used as the blank feature of each grid; The grid blank sequence features F 5 for: ,in, For the t The grid blank feature of the corresponding grid at each moment, For the t The number of white pixels in the grid corresponds to each moment. For the t The total number of pixels in the grid corresponds to each moment.

8. The calligraphy work complexity evaluation method based on layout aesthetics according to claim 1, It is characterized in that The state transfer matrix is ​​constructed by the state transfer probability distribution P for: ,in, The Markov chain value is m Status to n The transition probability value of the state, m =1, 2, 3 represent decreasing, stable and increasing states respectively; n =1, 2, 3 represent decreasing, stable and increasing states respectively; When the size complexity index is obtained by using the complexity index acquisition method, the Markov chain value is each character size feature in the size sequence feature; When the shape complexity index is obtained by using the complexity index acquisition method, the Markov chain value is each character shape feature in the shape sequence feature; When the distance complexity index is obtained by using the complexity index acquisition method, the Markov chain value is each character distance feature in the distance sequence feature; When the slope complexity index is obtained by using the complexity index acquisition method, the Markov chain value is the slope feature of each character in the slope sequence feature; When the complexity index obtaining method is used to obtain the grid blank complexity index, the Markov chain value is each grid blank feature in the grid blank sequence feature.

9. The calligraphy work complexity evaluation method based on layout aesthetics according to claim 8, It is characterized in that The Markov chain stationary distribution equation is used to solve the state transfer matrix to obtain the stationary distribution probabilities of the decreasing, stable and increasing states of the size sequence characteristics respectively. The Markov chain stationary distribution equation is: ,in, is the size sequence feature n The probability of a stationary distribution under a state is t The index of the time.

Images