An AI-based automatic calibration method and system for cultural and creative patterns

Abstract

This invention discloses an automatic calibration method and system for cultural and creative patterns based on artificial intelligence, relating to the fields of image processing and computer vision. The method includes: extracting deep visual features of the ice cream cultural and creative pattern to be calibrated and a target standard template using artificial intelligence technology, generating a multi-dimensional feature deviation vector in semantic structure and style features; performing co-adaptation of the pattern's content structure and visual style based on this vector to generate a pre-calibrated pattern; evaluating its semantic consistency and aesthetic quality score with the target template; calculating geometric calibration parameters and color calibration coefficients based on the comparison results of the scores and preset thresholds; and finally performing pixel-level transformation on the pattern to output the automatically calibrated ice cream cultural and creative pattern. This invention can effectively improve the consistency, aesthetics, and style uniformity of the pattern and template, significantly improving the efficiency and quality of cultural and creative pattern design.

Description

Technical Field

[0001] This invention relates to the fields of image processing, computer vision and artificial intelligence, and in particular to an automatic calibration method for cultural and creative patterns based on artificial intelligence. Background Technology

[0002] With the rapid development of the cultural and creative industries, the appearance design of fast-moving consumer goods such as ice cream has become a key factor in brand differentiation. Traditional cultural and creative graphic design relies heavily on manual adjustments, resulting in low efficiency, poor consistency, and inconsistent styles. In particular, the process of matching patterns with standard templates requires repeated adjustments to geometric structures and color styles, which is time-consuming and makes it difficult to guarantee quality.

[0003] While some image alignment and style transfer methods exist in the existing technology, they mostly focus on a single task, lack comprehensive consideration of semantic structure and visual style, and lack effective quality assessment and feedback mechanisms, making it difficult to meet the high-precision and high-aesthetic calibration requirements of cultural and creative patterns.

[0004] Therefore, there is an urgent need for a technical solution that can automate and intelligently complete pattern calibration while ensuring semantic consistency and aesthetic quality. Summary of the Invention

[0005] An artificial intelligence-based automatic calibration method and system for cultural and creative patterns, comprising:

[0006] S1. Use artificial intelligence technology to extract deep visual features of the ice cream cultural and creative patterns to be calibrated and the target standard template, and generate a multi-dimensional feature deviation vector in semantic structure and style features by comparison.

[0007] S2. Based on the multi-dimensional feature deviation vector, the pattern is adapted to the content structure and visual style to generate a pre-calibrated cultural and creative pattern.

[0008] S3. Evaluate the semantic consistency between the pattern after preliminary calibration and the target template, and calculate its aesthetic quality score;

[0009] S4. Compare the semantic consistency and aesthetic quality scores with preset thresholds to calculate the final geometric calibration parameters and color calibration coefficients.

[0010] S5. Based on the geometric calibration parameters and color calibration coefficients, perform the final pixel-level transformation on the pre-calibrated pattern and output the automatically calibrated ice cream-themed cultural and creative pattern.

[0011] The above-described AI-based automatic calibration method for cultural and creative patterns involves using AI technology to extract deep visual features of the ice cream cultural and creative pattern to be calibrated and the target standard template, and generating a multi-dimensional feature deviation vector in semantic structure and style features through comparison. The method includes the following sub-steps:

[0012] A deep convolutional neural network is used to extract the semantic and stylistic features of the ice cream cultural and creative pattern to be calibrated and the target standard template, respectively, and generate semantic feature vectors and stylistic feature vectors.

[0013] Calculate the cosine similarity between semantic feature vectors and the Euclidean distance between style feature vectors, and generate a multidimensional feature bias vector based on similarity and distance;

[0014] Based on the preset feature importance weights, the multidimensional feature deviation vectors are weighted and fused to obtain the final multidimensional feature deviation vector.

[0015] The above-described AI-based automatic calibration method for cultural and creative patterns, which involves co-adapting the content structure and visual style of the pattern based on a multi-dimensional feature deviation vector to generate a pre-calibrated cultural and creative pattern, includes the following sub-steps:

[0016] Based on the semantic deviation component in the multidimensional feature deviation vector, the content structure of the pattern to be calibrated is adjusted, including contour alignment and key point matching, so that it is consistent with the semantic structure of the target template.

[0017] Based on the style deviation component in the multidimensional feature deviation vector, adjust the visual style of the pattern to be calibrated, including color distribution and texture characteristics, to make it consistent with the style features of the target template.

[0018] By simultaneously adjusting the content structure and visual style, an image fusion algorithm is used to generate preliminarily calibrated cultural and creative patterns.

[0019] The artificial intelligence-based automatic calibration method for cultural and creative patterns, as described above, includes the following sub-steps: evaluating the semantic consistency between the pre-calibrated pattern and the target template, and calculating its aesthetic quality score.

[0020] Calculate the Euclidean distance between the pattern and the target template in the semantic feature space after preliminary calibration, and use it as the degree of semantic consistency;

[0021] The aesthetic features of the pattern after preliminary calibration are extracted, including color harmony, compositional balance and texture richness, and an aesthetic quality score is calculated based on the preset aesthetic evaluation criteria.

[0022] An overall quality assessment index is generated based on semantic consistency and aesthetic quality scores.

[0023] The artificial intelligence-based automatic calibration method for cultural and creative patterns, as described above, generates an overall quality assessment index based on semantic consistency and aesthetic quality scores, including the following sub-steps:

[0024] The overall quality score is calculated based on a weighted sum of semantic consistency and aesthetic quality scores.

[0025] If the overall quality score is lower than the preset quality standard, the feature extraction or collaborative adaptation parameters will be automatically adjusted.

[0026] The feature weights and adaptation parameters are calibrated a second time through an iterative optimization algorithm until the output pattern meets the quality requirements.

[0027] The above-described AI-based automatic calibration method for cultural and creative patterns, which compares semantic consistency and aesthetic quality scores with preset thresholds to calculate the final geometric calibration parameters and color calibration coefficients, includes the following sub-steps:

[0028] The semantic consistency level is compared with a preset semantic threshold. If it is lower than the threshold, the geometric transformation parameters, including scaling factor, rotation angle and translation vector, are calculated.

[0029] The aesthetic quality score is compared with the preset aesthetic threshold. If it is lower than the threshold, the color adjustment coefficient is calculated, including brightness gain, contrast factor and saturation adjustment value.

[0030] Based on the comparison results, geometric calibration parameters and color calibration coefficients are generated.

[0031] The above-described AI-based automatic calibration method for cultural and creative patterns includes the following sub-steps: Based on geometric calibration parameters and color calibration coefficients, the pre-calibrated pattern undergoes a final pixel-level transformation to output an automatically calibrated ice cream-themed cultural and creative pattern.

[0032] Based on the geometric calibration parameters, the pre-calibrated pattern is subjected to geometric transformations, including scaling, rotation, and translation operations, to accurately align with the target template;

[0033] Based on the color calibration coefficient, the color of the geometrically transformed pattern is adjusted, including pixel-level correction of brightness, contrast and saturation.

[0034] The adjusted pattern is smoothed at the pixel level to eliminate edge distortion and color banding, and the resulting ice cream-themed cultural and creative pattern is automatically calibrated.

[0035] An AI-based automatic calibration system for cultural and creative patterns, comprising:

[0036] Feature extraction and deviation calculation module: Utilizes artificial intelligence technology to extract deep visual features of the ice cream cultural and creative pattern to be calibrated and the target standard template, and generates a multi-dimensional feature deviation vector in semantic structure and style features through comparison;

[0037] Preliminary calibration module: Based on multi-dimensional feature deviation vectors, the content structure and visual style of the pattern are adapted collaboratively to generate a preliminarily calibrated cultural and creative pattern;

[0038] Quality assessment and parameter calculation module: assesses the semantic consistency between the pattern and the target template after preliminary calibration and calculates its aesthetic quality score; compares the semantic consistency and aesthetic quality score with preset thresholds to calculate the final geometric calibration parameters and color calibration coefficients;

[0039] Final calibration output module: Based on the geometric calibration parameters and color calibration coefficients, the module performs a final pixel-level transformation on the pre-calibrated pattern and outputs the automatically calibrated ice cream-themed cultural and creative pattern.

[0040] A computer storage medium, characterized in that it comprises: at least one memory and at least one processor;

[0041] Memory, used to store one or more program instructions;

[0042] A processor for running one or more program instructions to execute an AI-based automatic calibration method for cultural and creative patterns as described above.

[0043] The beneficial effects achieved by this invention are as follows:

[0044] By using deep learning and attention mechanisms to accurately extract the semantic and stylistic features of patterns, high-precision deviation analysis and intelligent adaptation of content structure are achieved. Combining dual constraints of content and style with generative adversarial training, the consistency, visual naturalness, and stylistic unity of patterns and templates are effectively improved. A multi-dimensional quality assessment and feedback mechanism is introduced to ensure that the output patterns have both high semantic consistency and aesthetic quality. Finally, through parametric geometric transformation and color calibration, pixel-level precise adjustment is achieved, significantly improving the design efficiency, quality stability, and automation level of cultural and creative patterns, making it suitable for large-scale applications in the fast-moving consumer goods sector, such as ice cream packaging. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0046] Figure 1 This is a flowchart of an automatic calibration method for cultural and creative patterns based on artificial intelligence, provided in an embodiment of this application.

[0047] Figure 2 This is a schematic diagram of an artificial intelligence-based automatic calibration system for cultural and creative patterns provided in an embodiment of this application. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] Example 1

[0050] like Figure 1 As shown in the figure, an embodiment of this application provides an automatic calibration method for cultural and creative patterns based on artificial intelligence, comprising:

[0051] Step S1: Use artificial intelligence technology to extract deep visual features of the ice cream cultural and creative pattern to be calibrated and the target standard template, and generate a multi-dimensional feature deviation vector in semantic structure and style features by comparison.

[0052] Specifically, a deep convolutional neural network is used to extract semantic and stylistic features from the ice cream cultural and creative patterns to be calibrated and the target standard template, generating semantic feature vectors and stylistic feature vectors. The cosine similarity between semantic feature vectors and the Euclidean distance between stylistic feature vectors are calculated. A multidimensional feature deviation vector is generated based on the similarity and distance. The multidimensional feature deviation vector is then weighted and fused according to preset feature importance weights to obtain the final multidimensional feature deviation vector. The process includes the following sub-steps:

[0053] Step S11: Use a deep convolutional neural network to extract the semantic features and style features of the ice cream cultural and creative pattern to be calibrated and the target standard template respectively, and generate semantic feature vector and style feature vector;

[0054] Image preprocessing operations are performed to standardize the ice cream-themed design and the target standard template, including adjusting image size, normalizing brightness and contrast, and removing noise interference, thereby improving the robustness of feature extraction. Feature extraction algorithms are then applied to capture semantic features from the preprocessed images. These features reflect the core content and structural information of the design, resulting in a high-dimensional semantic feature vector. Simultaneously, for style features, statistical characteristics related to style are extracted by analyzing the color distribution, brushstroke style, and consistency of artistic elements, forming a style feature vector.

[0055] Step S12: Calculate the cosine similarity between semantic feature vectors and the Euclidean distance between style feature vectors, and generate a multidimensional feature deviation vector based on similarity and distance;

[0056] The generated semantic feature vectors are normalized to eliminate the influence of dimensions, and then the cosine similarity between the vectors is calculated using the following formula:

[0057] in, Indicates the first Cosine similarity between vectors at each semantic layer; Indicates the image to be calibrated at the 1st... Semantic feature vectors of each semantic layer; The standard template is in the first... Semantic feature vectors of each semantic layer; Indicates the transpose operation; This represents a hyperparameter used to control the contribution weight of the "Gaussian kernel term" to the final similarity. This represents the variance of the Gaussian function.

[0058] This similarity score is used to assess semantic structural consistency; a higher value indicates that the pattern is closer to the standard template at the content level. For style feature vectors, style differences are quantified by calculating the Euclidean distance; a smaller distance indicates greater style similarity. The Euclidean distance of style features is expressed by the following formula:

[0059] in, Indicates the Euclidean distance of style characteristics; Indicates the total number of style layers; Indicates the feature dimension index; Indicates the first The weights of each style layer; This represents the style feature vector of the object to be calibrated at the k-th style layer; The standard template is in the first... Style feature vectors of each style layer; Indicates the first Standard deviation estimation of 3D features.

[0060] This reflects the deviations in the artistic expression of the patterns. Based on these calculations, similarity and distance values ​​are combined into a preliminary multi-dimensional feature deviation vector, where each dimension corresponds to a feature metric. To enhance the discriminative power of the vector, a dynamic threshold mechanism is introduced. A reasonable range is set based on historical data or domain knowledge to filter out irrelevant noise and ensure that the deviation vector focuses on key differences. Furthermore, feature change trends are monitored in real time, and calculation parameters are adjusted to adapt to the variability of different pattern types, improving the adaptability of deviation generation.

[0061] Step S13: According to the preset feature importance weights, the multidimensional feature deviation vector is weighted and fused to obtain the final multidimensional feature deviation vector;

[0062] Based on the application scenario requirements, feature importance weights are predefined. A weighted fusion algorithm is used to process the initial multi-dimensional feature deviation vector. Each dimension is scaled and integrated according to the weight coefficients to generate the final deviation vector. An adaptive adjustment mechanism is introduced during the process. By evaluating the distribution characteristics of the deviation vector in real time, the weight allocation is dynamically optimized to avoid overfitting or amplification of deviations. Simultaneously, a quality check is established to verify the stability and effectiveness of the fused vector. If abnormal fluctuations are detected, the weight parameters are adjusted until the vector meets the accuracy and consistency requirements of the calibration, ensuring that the output results can reliably guide subsequent calibration operations.

[0063] Step S2: Based on the multi-dimensional feature deviation vector, the content structure and visual style of the pattern are collaboratively adapted to generate a preliminarily calibrated cultural and creative pattern.

[0064] Specifically, based on the semantic deviation component in the multidimensional feature deviation vector, the content structure of the pattern to be calibrated is optimized, and consistency with the semantic structure of the target template is achieved through contour alignment and key point matching. Simultaneously, the visual style of the pattern is adjusted according to the style deviation component, encompassing the adaptation of color distribution and texture characteristics to ensure harmony with the style features of the target template. Building upon this, through the coordinated adjustment of content structure and visual style, an image fusion algorithm is used to generate a preliminarily calibrated cultural and creative pattern, including the following sub-steps:

[0065] Step S21: Adjust the content structure of the pattern to be calibrated, including contour alignment and key point matching, according to the semantic deviation component in the multidimensional feature deviation vector, so that it is consistent with the semantic structure of the target template.

[0066] The contour information of the pattern to be calibrated and the target template are obtained through a contour extraction algorithm, and the shape deviation between the two is calculated using a formula:

[0067] Identify key contour points such as corner points or curvature extrema; based on the correspondence of key points, use geometric transformation algorithms such as affine transformation or perspective transformation to spatially adjust the pattern to be calibrated, ensuring that the contour boundary is aligned with the target template; at the same time, evaluate the consistency of feature points through key point matching formulas, iteratively optimize transformation parameters, eliminate semantic structure deviations, and achieve accurate adaptation of content structure.

[0068] Step S22: Adjust the visual style of the pattern to be calibrated, including color distribution and texture characteristics, according to the style deviation component in the multidimensional feature deviation vector, so that it is consistent with the style characteristics of the target template.

[0069] The color distribution difference between the pattern to be calibrated and the target template is calculated using a color histogram analysis formula. The hue, saturation, and brightness of the pattern are adjusted using color mapping technology to make its color statistical characteristics converge with those of the target template. For texture characteristics, the local texture patterns of the pattern are analyzed, such as calculating texture contrast and uniformity based on the gray-level co-occurrence matrix or Gabor filter. Texture synthesis or adaptation algorithms are then used to smoothly adjust texture details, ensuring a natural transition and consistency in visual style.

[0070] Step S23: Through the synchronous adjustment of content structure and visual style, an image fusion algorithm is used to generate a preliminarily calibrated cultural and creative pattern;

[0071] After completing contour alignment and color / texture adaptation, multi-resolution image fusion techniques such as Laplacian pyramid fusion or wavelet transform fusion are used to seamlessly integrate the adjusted content structure with the visual style layer; the contribution of each pixel is calculated using a formula.

[0072] in, Indicates the coordinates of the fused image Pixel value at; Represents the contribution weight at the pixel level; This indicates the image to be calibrated in coordinates. Pixel value at; Indicates "standard template / reference image" "The contribution ratio at this pixel location;" Indicates the standard template in coordinates The pixel value at that location.

[0073] Regions with clear semantic structures are prioritized for preservation, while style transition boundaries are smoothed to avoid artifacts or distortion. Finally, after generating the fusion result, quality evaluation metrics such as structural similarity or peak signal-to-noise ratio are used for preliminary verification to ensure that the calibration pattern is visually coherent and conforms to the target specifications.

[0074] Step S3: Evaluate the semantic consistency between the pattern after preliminary calibration and the target template, and calculate its aesthetic quality score;

[0075] Specifically, the semantic consistency is quantified by calculating the Euclidean distance between the pre-calibrated pattern and the target template in the semantic feature space. Simultaneously, aesthetic features of the pre-calibrated pattern are extracted, including color harmony, compositional balance, and texture richness. An aesthetic quality score is calculated based on preset aesthetic evaluation criteria. Finally, the overall quality assessment index is generated by combining the semantic consistency and aesthetic quality scores, including the following sub-steps:

[0076] Step S31: Calculate the Euclidean distance between the pattern after preliminary calibration and the target template in the semantic feature space, as the degree of semantic consistency;

[0077] A feature extraction algorithm is used to generate semantic feature vectors for the pre-calibrated pattern and target template. A pre-trained semantic encoder converts the pattern into a high-dimensional feature representation, ensuring that the feature vectors can capture the deep semantic information of the pattern, such as topic consistency and content relevance. The Euclidean distance is calculated using the following formula:

[0078] in, Euclidean distance represents semantic consistency; This represents the semantic feature vector to be processed; A semantic feature vector representing the target; Representing vectors transpose; Represents the feature weight matrix; This represents the total dimension of the semantic feature vector; Indicates the first Weights of dimensional features; Represents the vector to be processed With the target vector In the Differences in dimensional features.

[0079] A smaller value indicates higher semantic consistency. To ensure computational accuracy, feature normalization is introduced to eliminate the impact of scale differences, and distance change trends are recorded in real time to provide data support for subsequent adjustments.

[0080] Step S32: Extract the aesthetic features of the pattern after preliminary calibration, including color harmony, compositional balance and texture richness, and calculate the aesthetic quality score based on the preset aesthetic evaluation criteria.

[0081] Aesthetic features of the pre-calibrated pattern are extracted, including color harmony, compositional balance, and texture richness, and an aesthetic quality score is calculated based on preset aesthetic evaluation criteria. For color harmony, a color distribution analysis algorithm is used to calculate the uniformity and complementarity of the main colors in the pattern. The coordination of color combinations is evaluated using parameters such as hue, saturation, and brightness to avoid conflicting or jarring areas. Compositional balance is assessed by calculating the spatial symmetry and center-of-gravity position of elements in the pattern, using the offset between the geometric center and the visual center of gravity as an indicator to ensure layout stability. Texture richness is calculated by analyzing the local texture details and repetition patterns of the pattern, using statistical methods to calculate the texture complexity index. Based on these features, an aesthetic quality score is calculated using a formula:

[0082]

[0083] in, This indicates the overall aesthetic quality score; It indicates color harmony and assesses the coordination and aesthetics of color matching in a pattern; It indicates the compositional balance and assesses the degree of visual balance in the arrangement of elements in the pattern; It indicates texture richness, assessing the complexity and richness of the texture details in a pattern.

[0084] The weights of each feature are dynamically adjusted according to preset standards, and the scoring results reflect the overall aesthetic level of the pattern.

[0085] Step S33: Generate an overall quality assessment index based on the degree of semantic consistency and aesthetic quality score;

[0086] Step S331: Calculate the overall quality score based on the weighted sum of semantic consistency and aesthetic quality scores;

[0087] By defining weighting coefficients, the semantic consistency distance value and aesthetic score are combined to ensure that the contributions of the two are balanced, and the higher the score, the better the pattern quality.

[0088] Step S332: If the overall quality score is lower than the preset quality standard, the feature extraction or collaborative adaptation parameters will be automatically adjusted.

[0089] Step S333: Perform secondary calibration of feature weights and adaptation parameters through iterative optimization algorithm until the output pattern meets the quality requirements;

[0090] The feature weights and adaptation parameters are calibrated a second time by iterative optimization algorithm. The optimal parameter combination is searched by gradient descent method and historical evaluation data is combined to accelerate convergence until the output pattern meets the quality requirements.

[0091] Step S4: Compare the semantic consistency and aesthetic quality scores with preset thresholds to calculate the final geometric calibration parameters and color calibration coefficients;

[0092] Specifically, the semantic consistency level is compared with a preset semantic threshold. If it is lower than the threshold, geometric transformation parameters, including scaling factors, rotation angles, and translation vectors, are calculated. Simultaneously, the aesthetic quality score is compared with a preset aesthetic threshold. If it is lower than the threshold, color adjustment coefficients are calculated, involving brightness gain, contrast factor, and saturation adjustment values. Based on the comparison results, geometric calibration parameters and color calibration coefficients are generated comprehensively, including the following sub-steps:

[0093] Step S41: Compare the semantic consistency level with the preset semantic threshold. If it is lower than the threshold, calculate the geometric transformation parameters, including scaling factor, rotation angle and translation vector.

[0094] Key semantic feature points are extracted from the image and matched with a reference standard with high precision. Then, based on the spatial distribution of the matched point pairs, the optimal transformation parameter—the scaling factor—is calculated. This scaling factor is determined by the ratio of the bounding box size formed by the feature points and is expressed by the following formula:

[0095]

[0096] in, Indicates the scaling factor; Indicates the eigenvalues ​​on the reference side; This represents the feature value of the side to be scaled; Represents the reference matrix; Represents the matrix to be scaled; This represents the Frobenius norm.

[0097] The rotation angle is calculated by the angle between the lines connecting the feature points along the main direction, and is expressed by the following formula:

[0098]

[0099] in, Indicates the rotation angle; Represents the two-dimensional arctangent function; Representing the eigenvector of Quantity, Quantity; Representing the eigenvector of Quantity, Quantity; Indicates the regulating factor; Represents the reference eigenvector; This represents the eigenvector to be rotated.

[0100] The translation vector is derived from the centroid offset of the feature point set, and is expressed by the following formula:

[0101]

[0102] in, Represents the translation vector; Represents the average position vector of the reference point; Indicates the scaling factor; Indicates rotation angle The cosine and sine values; This represents the average position vector of the point to be transformed.

[0103] The final result is a set of geometric transformation parameters that can accurately correct spatial relationships.

[0104] Step S42: Compare the aesthetic quality score with the preset aesthetic threshold. If it is lower than the threshold, calculate the color adjustment coefficient, including brightness gain, contrast factor and saturation adjustment value.

[0105] By analyzing the statistical distribution of the image in the HSL color space and combining it with a visual perception model to dynamically calculate and adjust parameters—the brightness gain is determined based on the deviation between the centroid of the histogram distribution and the ideal brightness-dark relationship, and is expressed by the following formula:

[0106]

[0107] in, Indicates brightness gain; Represents the natural exponential function; This represents the average value of ideal brightness. This represents the average current brightness. Indicates the tuning parameters; Represents characteristic quantities related to brightness; Indicates the tuning parameters; Represents the hyperbolic tangent function; This represents the standard deviation of the current brightness.

[0108] The contrast factor is adaptively generated by analyzing the intensity dynamic range of high and low frequency components, and is expressed by the following formula:

[0109]

[0110] in, Indicates the contrast factor; This represents the standard deviation of high-frequency components under ideal conditions; This represents the standard deviation of the current high-frequency components; Indicates the tuning parameters; This represents the standard deviation of the current brightness. This represents the average current brightness. Represents a logarithmic function.

[0111] The saturation adjustment value is optimized based on the color distribution dispersion and visual comfort model, and is expressed by the following formula:

[0112]

[0113] in, This indicates the saturation adjustment value; The weighting parameter represents the mean term; The mean value representing the ideal saturation; This represents the mean of the current saturation level; The weighting parameter represents the standard deviation term; The standard deviation represents the ideal saturation. This represents the standard deviation of the current saturation level. The weighting parameters represent the energy term; Energy representing ideal saturation; Energy representing the current saturation level.

[0114] The final result is a color correction factor that enhances visual appeal while maintaining a natural feel.

[0115] Step S43: Based on the comparison results, generate geometric calibration parameters and color calibration coefficients.

[0116] Based on the comparison results of steps S41 and S42, geometric calibration parameters and color calibration coefficients are generated comprehensively. If both semantic consistency and aesthetic quality scores are below the threshold, geometric and color calibrations are applied simultaneously; if only one is below the threshold, that item is processed first to optimize efficiency. The generation process first integrates geometric transformation parameters and color adjustment coefficients, performs parameter normalization, and ensures consistency among the coefficients. A weighted fusion algorithm combines geometric and color factors to avoid parameter conflicts or over-adjustment, ultimately outputting a complete set of calibration parameters.

[0117] Step S5: Based on the geometric calibration parameters and color calibration coefficients, perform the final pixel-level transformation on the pre-calibrated pattern to output the automatically calibrated ice cream-themed cultural and creative pattern.

[0118] Specifically, the process involves performing a geometric transformation on the initially calibrated pattern based on geometric calibration parameters, achieving precise alignment with the target template through scaling, rotation, and translation operations; adjusting the color of the transformed pattern according to color calibration coefficients, involving pixel-level corrections to brightness, contrast, and saturation; and performing pixel-level smooth blending on the adjusted pattern to eliminate edge distortion and color banding, outputting an automatically calibrated ice cream-themed creative pattern. This includes the following sub-steps:

[0119] Step S51: Based on the geometric calibration parameters, perform geometric transformations on the pre-calibrated pattern, including scaling, rotation, and translation operations, to accurately align with the target template;

[0120] The pattern is scaled to match the target template specifications; a rotation transformation is applied by calculating the rotation matrix to correct directional deviations; and a displacement operation is performed based on the translation vector to ensure precise spatial positioning of the pattern. The entire process dynamically adjusts the transformation matrix based on preset calibration parameters, guaranteeing the accuracy and efficiency of geometric alignment.

[0121] Step S52: Adjust the color of the geometrically transformed pattern according to the color calibration coefficient, including pixel-level correction of brightness, contrast and saturation.

[0122] The overall brightness distribution of the image is optimized by adjusting the brightness gain coefficient to avoid overexposure or underexposure; a contrast stretching function is applied to enhance the dynamic range of the image and improve detail; and a saturation correction algorithm is used to adjust color vibrancy so that the pattern colors meet preset standards. All adjustments are based on pixel-level calculations to ensure color consistency and natural transitions while avoiding the introduction of noise or distortion.

[0123] Step S53: Perform pixel-level smoothing and blending on the adjusted pattern to eliminate edge distortion and color banding, and output the ice cream cultural and creative pattern that has completed automatic calibration.

[0124] Gaussian filtering or bilateral filtering algorithms are used to smooth the edges of the pattern, reducing jagged edges and uneven transitions. Color interpolation techniques are used to process color gradient areas to ensure natural blending. The final output is an automatically calibrated ice cream-themed pattern, ensuring complete image quality and visual flawlessness. Throughout the process, real-time monitoring of the blending effect is emphasized, with fine-tuning as necessary to achieve high-quality output results.

[0125] Example 2

[0126] like Figure 2 As shown, Embodiment 2 of this application provides an automatic calibration system for cultural and creative patterns based on artificial intelligence, including:

[0127] Feature Extraction and Bias Calculation 21: Utilizing artificial intelligence technology, deep visual features of the ice cream cultural and creative patterns to be calibrated and the target standard template are extracted. A multi-dimensional feature bias vector in semantic structure and style features is generated through comparison, including the following sub-modules:

[0128] Feature extraction submodule 211: Uses a deep convolutional neural network to extract the semantic features and style features of the ice cream cultural and creative pattern to be calibrated and the target standard template respectively, and generates semantic feature vector and style feature vector;

[0129] Deviation calculation submodule 212: Calculates the cosine similarity between semantic feature vectors and the Euclidean distance between style feature vectors, and generates a multidimensional feature deviation vector based on similarity and distance;

[0130] Weighted fusion submodule 213: Based on the preset feature importance weights, the multidimensional feature deviation vector is weighted and fused to obtain the final multidimensional feature deviation vector;

[0131] Preliminary calibration module 22: Based on the multi-dimensional feature deviation vector, the content structure and visual style of the pattern are collaboratively adapted to generate the preliminarily calibrated cultural and creative pattern, including the following sub-modules:

[0132] Content structure calibration submodule 221: Adjusts the content structure of the pattern to be calibrated, including contour alignment and key point matching, based on the semantic deviation component in the multidimensional feature deviation vector, so that it is consistent with the semantic structure of the target template.

[0133] Structural calibration submodule 222: Adjusts the visual style of the pattern to be calibrated, including color distribution and texture characteristics, according to the style deviation component in the multidimensional feature deviation vector, so that it is consistent with the style characteristics of the target template;

[0134] Fusion Generation Submodule 223: Through synchronous adjustment of content structure and visual style, an image fusion algorithm is used to generate cultural and creative patterns after preliminary calibration;

[0135] Quality assessment and parameter calculation module 23: assesses the semantic consistency between the pattern and the target template after preliminary calibration and calculates its aesthetic quality score; compares the semantic consistency and aesthetic quality score with preset thresholds to calculate the final geometric calibration parameters and color calibration coefficients, including the following sub-modules:

[0136] Semantic evaluation submodule 231: Calculates the Euclidean distance between the pattern after preliminary calibration and the target template in the semantic feature space, as the degree of semantic consistency;

[0137] Aesthetic evaluation submodule 232: Extracts the aesthetic features of the pattern after preliminary calibration, including color harmony, compositional balance and texture richness, and calculates the aesthetic quality score based on the preset aesthetic evaluation criteria;

[0138] Comprehensive evaluation submodule 233: Generates overall quality evaluation indicators based on semantic consistency and aesthetic quality scores;

[0139] Geometric parameter calculation submodule 234: compares the semantic consistency degree with a preset semantic threshold. If it is lower than the threshold, it calculates the geometric transformation parameters, including scaling factor, rotation angle and translation vector.

[0140] Color coefficient calculation submodule 235: Compares the aesthetic quality score with the preset aesthetic threshold. If it is lower than the threshold, it calculates the color adjustment coefficient, including brightness gain, contrast factor and saturation adjustment value.

[0141] Calibration parameter integration submodule 236: Based on the comparison results, it integrates geometric calibration parameters and color calibration coefficients;

[0142] Final calibration output module 24: Based on the geometric calibration parameters and color calibration coefficients, performs final pixel-level transformation on the pre-calibrated pattern, outputting the automatically calibrated ice cream-themed cultural and creative pattern, including the following sub-modules:

[0143] Geometric calibration submodule 241: Based on the geometric calibration parameters, perform geometric transformations on the pre-calibrated pattern, including scaling, rotation and translation operations, to accurately align with the target template;

[0144] Color calibration submodule 242: Adjusts the color of the geometrically transformed pattern according to the color calibration coefficient, including pixel-level correction of brightness, contrast and saturation;

[0145] Output optimization submodule 243: Performs pixel-level smooth fusion on the adjusted pattern to eliminate edge distortion and color banding, and outputs an automatically calibrated ice cream-themed cultural and creative pattern.

[0146] Corresponding to the above embodiments, the present invention provides a computer storage medium, including: at least one memory and at least one processor;

[0147] The memory is used to store one or more program instructions;

[0148] A processor for running one or more program instructions to execute an artificial intelligence-based automatic calibration method for cultural and creative patterns;

[0149] Corresponding to the above embodiments, this embodiment of the invention provides a computer-readable storage medium containing one or more program instructions, which are executed by a processor to provide an artificial intelligence-based automatic calibration method for cultural and creative patterns.

[0150] The embodiments disclosed in this invention provide a computer-readable storage medium storing computer program instructions. When the computer program instructions are executed on a computer, the computer performs the above-described artificial intelligence-based automatic calibration method for cultural and creative patterns.

[0151] In this embodiment of the invention, the processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0152] The various methods, steps, and logic diagrams disclosed in the embodiments of this invention can be implemented or executed. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The processor reads information from the storage medium and, in conjunction with its hardware, completes the steps of the above methods.

[0153] The storage medium can be memory, such as volatile memory or non-volatile memory, or may include both volatile and non-volatile memory.

[0154] Among them, non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory.

[0155] Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (Synchlink DRAM, SLDRAM), and direct memory bus RAM (DRRAM).

[0156] The storage media described in the embodiments of the present invention are intended to include, but are not limited to, these and any other suitable types of memory.

[0157] Those skilled in the art will recognize that, in one or more of the examples above, the functions described in this invention can be implemented using a combination of hardware and software. When applied as software, the corresponding functions can be stored in a computer-readable medium or transmitted as one or more instructions or code on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media include any medium that facilitates the transmission of computer programs from one place to another. Storage media can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0158] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, or improvements made on the basis of the technical solution of the present invention should be included within the scope of protection of the present invention.

Claims

1. An automatic calibration method for cultural and creative patterns based on artificial intelligence, characterized in that, include: S1. Use artificial intelligence technology to extract deep visual features of the ice cream cultural and creative patterns to be calibrated and the target standard template, and generate a multi-dimensional feature deviation vector in semantic structure and style features by comparison. S2. Based on the multi-dimensional feature deviation vector, the pattern is collaboratively adapted in terms of content structure and visual style to generate a pre-calibrated cultural and creative pattern. This includes the following sub-steps: Based on the semantic deviation component in the multidimensional feature deviation vector, the content structure of the pattern to be calibrated is adjusted, including contour alignment and key point matching, so that it is consistent with the semantic structure of the target template. Based on the style deviation component in the multidimensional feature deviation vector, adjust the visual style of the pattern to be calibrated, including color distribution and texture characteristics, to make it consistent with the style features of the target template. By simultaneously adjusting the content structure and visual style, an image fusion algorithm is used to generate preliminarily calibrated cultural and creative patterns. S3. Evaluate the semantic consistency between the pattern after preliminary calibration and the target template, and calculate its aesthetic quality score; S4. Compare the semantic consistency and aesthetic quality scores with preset thresholds to calculate the final geometric calibration parameters and color calibration coefficients. S5. Based on the geometric calibration parameters and color calibration coefficients, perform the final pixel-level transformation on the pre-calibrated pattern and output the automatically calibrated ice cream-themed cultural and creative pattern.

2. The method for automatic calibration of cultural and creative patterns based on artificial intelligence according to claim 1, characterized in that, Artificial intelligence technology is used to extract deep visual features of the ice cream cultural and creative patterns to be calibrated and the target standard template. By comparison, a multi-dimensional feature deviation vector in semantic structure and style features is generated, including the following sub-steps: A deep convolutional neural network is used to extract the semantic and stylistic features of the ice cream cultural and creative pattern to be calibrated and the target standard template, respectively, and generate semantic feature vectors and stylistic feature vectors. Calculate the cosine similarity between semantic feature vectors and the Euclidean distance between style feature vectors, and generate a multidimensional feature bias vector based on similarity and distance; Based on the preset feature importance weights, the multidimensional feature deviation vectors are weighted and fused to obtain the final multidimensional feature deviation vector.

3. The method for automatic calibration of cultural and creative patterns based on artificial intelligence according to claim 1, characterized in that, Assess the semantic consistency between the initial calibrated pattern and the target template, and calculate its aesthetic quality score, including the following sub-steps: Calculate the Euclidean distance between the pattern and the target template in the semantic feature space after preliminary calibration, and use it as the degree of semantic consistency; The aesthetic features of the pattern after preliminary calibration are extracted, including color harmony, compositional balance and texture richness, and an aesthetic quality score is calculated based on the preset aesthetic evaluation criteria. An overall quality assessment index is generated based on semantic consistency and aesthetic quality scores.

4. The method for automatic calibration of cultural and creative patterns based on artificial intelligence according to claim 3, characterized in that, Based on semantic consistency and aesthetic quality scores, an overall quality assessment index is generated, including the following sub-steps: The overall quality score is calculated based on a weighted sum of semantic consistency and aesthetic quality scores. If the overall quality score is lower than the preset quality standard, the feature extraction or collaborative adaptation parameters will be automatically adjusted. The feature weights and adaptation parameters are calibrated a second time through an iterative optimization algorithm until the output pattern meets the quality requirements.

5. The method for automatic calibration of cultural and creative patterns based on artificial intelligence according to claim 1, characterized in that, The semantic consistency and aesthetic quality scores are compared with preset thresholds to calculate the final geometric calibration parameters and color calibration coefficients, including the following sub-steps: The semantic consistency level is compared with a preset semantic threshold. If it is lower than the threshold, the geometric transformation parameters, including scaling factor, rotation angle and translation vector, are calculated. The aesthetic quality score is compared with the preset aesthetic threshold. If it is lower than the threshold, the color adjustment coefficient is calculated, including brightness gain, contrast factor and saturation adjustment value. Based on the comparison results, geometric calibration parameters and color calibration coefficients are generated.

6. The method for automatic calibration of cultural and creative patterns based on artificial intelligence according to claim 1, characterized in that, Based on the geometric calibration parameters and color calibration coefficients, the pre-calibrated pattern undergoes a final pixel-level transformation to output an automatically calibrated ice cream-themed cultural and creative pattern. This process includes the following sub-steps: Based on the geometric calibration parameters, the pre-calibrated pattern is subjected to geometric transformations, including scaling, rotation, and translation operations, to accurately align with the target template. Based on the color calibration coefficient, the color of the geometrically transformed pattern is adjusted, including pixel-level correction of brightness, contrast and saturation. The adjusted pattern is smoothed at the pixel level to eliminate edge distortion and color banding, and the resulting ice cream-themed cultural and creative pattern is automatically calibrated.

7. An automatic calibration system for cultural and creative patterns based on artificial intelligence, characterized in that, include: Feature extraction and deviation calculation module: Utilizes artificial intelligence technology to extract deep visual features of the ice cream cultural and creative pattern to be calibrated and the target standard template, and generates a multi-dimensional feature deviation vector in semantic structure and style features through comparison; Preliminary calibration module: Based on multi-dimensional feature deviation vectors, the content structure and visual style of the pattern are adapted collaboratively to generate a preliminarily calibrated cultural and creative pattern; The preliminary calibration module is specifically used to adjust the content structure of the pattern to be calibrated, including contour alignment and key point matching, according to the semantic deviation component in the multidimensional feature deviation vector, so that it is consistent with the semantic structure of the target template; and to adjust the visual style of the pattern to be calibrated, including color distribution and texture characteristics, according to the style deviation component in the multidimensional feature deviation vector, so that it is coordinated with the style characteristics of the target template. By simultaneously adjusting the content structure and visual style, an image fusion algorithm is used to generate preliminarily calibrated cultural and creative patterns. Quality assessment and parameter calculation module: assesses the semantic consistency between the pattern and the target template after initial calibration, and calculates its aesthetic quality score; The semantic consistency and aesthetic quality scores are compared with preset thresholds to calculate the final geometric calibration parameters and color calibration coefficients. Final calibration output module: Based on the geometric calibration parameters and color calibration coefficients, the module performs a final pixel-level transformation on the pre-calibrated pattern and outputs the automatically calibrated ice cream-themed cultural and creative pattern.

8. A computer storage medium, characterized in that, include: At least one memory and at least one processor; Memory, used to store one or more program instructions; A processor for running one or more program instructions to execute an AI-based automatic calibration method for cultural and creative patterns as described in any one of claims 1-6.

Images