A method and system for generating a stop-motion animation, and a storage medium
By identifying key sentences in the animation script to generate keyframe images, and using AI recognition and designer adjustments to create supplementary frame images in a loop, the problem of low efficiency in stop-motion animation generation is solved, enabling the rapid generation of efficient stop-motion animation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ANHUI UNIVERSITY OF ARCHITECTURE
- Filing Date
- 2025-06-03
- Publication Date
- 2026-06-19
Smart Images
Figure CN120635257B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of stop-motion animation generation technology, specifically a method, system, and storage medium for generating stop-motion animation. Background Technology
[0002] Stop-motion animation is a form of animation that creates the illusion of motion by capturing still images frame by frame and playing them in sequence. Although this method is simple, it requires capturing a large number of still images, which is quite labor-intensive. Moreover, the shooting process is easily affected by the environment. Under the current technological background, many images can be directly synthesized, which makes the generation process of stop-motion animation more convenient. How to provide a fast stop-motion animation generation solution when the accuracy requirements are not high, so as to improve the production efficiency of stop-motion animation, is the technical problem that this invention aims to solve. Summary of the Invention
[0003] The purpose of this invention is to provide a method, system, and storage medium for generating stop-motion animation, in order to solve the problems mentioned in the background art.
[0004] To achieve the above objectives, the present invention provides the following technical solution:
[0005] A method, system, and storage medium for generating stop-motion animation, the method comprising:
[0006] The system receives animation scripts uploaded by designers, performs text recognition on the sentences in the animation scripts in turn, and obtains the noun groups of each sentence;
[0007] Based on the sentence order, noun groups are statistically analyzed to identify key sentences.
[0008] Read the noun phrases of key statements, read the animation components based on the noun phrases, and generate keyframe images based on the animation components;
[0009] Preserve the statement order, count keyframe images, select adjacent keyframe images in sequence, create and display supplementary frame images based on adjacent keyframe images, and receive adjustment information input by the designer;
[0010] The adjusted supplementary frame images are used as keyframe images. All keyframe images are counted, and the understanding is generated based on the keyframe images identified by AI.
[0011] The adjustment range of the adjustment information is obtained, the number of cyclic supplements is determined according to the adjustment range, and the process of creating supplementary frame images is executed cyclically until the comprehension reaches a preset comprehension threshold.
[0012] As a further aspect of the present invention: the step of identifying key sentences by statistically analyzing noun groups based on sentence order includes:
[0013] Number the statements sequentially and read the noun phrases in each statement in turn;
[0014] Each noun in the noun group is input into a word vector conversion model to obtain a word vector for each noun;
[0015] Calculate the word vector for each noun to obtain the vector group corresponding to the noun group;
[0016] Input the vector group into the preset numerical transformation model to obtain the feature values of the vector group;
[0017] Using the number as the independent variable and the eigenvalue as the dependent variable, coordinate points are constructed, and the eigenvalue variation function is fitted.
[0018] Obtain the derivative of the eigenvalue change function, mark the function segment whose derivative reaches a preset threshold, query the number of the marked function segment, and use the statement corresponding to the number as the key statement.
[0019] As a further aspect of the present invention: the steps of reading noun groups of key statements, reading animation components based on noun groups, and generating keyframe images based on animation components include:
[0020] Read the noun phrases in the key sentences;
[0021] For each noun, a component is matched from a preset component library and used as an animation component;
[0022] Query the type of animation component, combine animation components based on type, and generate keyframe images;
[0023] Displays keyframe images and receives adjustment information input by the designer.
[0024] As a further aspect of the present invention: the steps of preserving the statement order, statistically analyzing keyframe images, sequentially selecting adjacent keyframe images, creating and displaying supplementary frame images based on adjacent keyframe images, and receiving adjustment information input by the designer include:
[0025] Preserve statement order and analyze keyframe images;
[0026] Select adjacent keyframe images sequentially, compare the animation components of adjacent keyframe images, and determine the change vector of each animation component;
[0027] The supplementary positions of each animation component are determined based on the change vector, the supplementary positions of all animation components are counted, and a supplementary frame image is constructed.
[0028] Displays supplementary frame images and receives adjustment information input by the designer;
[0029] The process of determining the change vector for each animation component includes:
[0030] For animation components that appear in the previous keyframe image but not in the subsequent keyframe image, generate a change vector pointing from the animation component to outside the image frame;
[0031] For an animated component that appears in the previous keyframe image and in the next keyframe image, generate a change vector from the previous position to the next position of the animated component.
[0032] For an animated component that does not appear in the previous keyframe but appears in the next keyframe, generate a change vector pointing from outside the frame to the position of the animated component.
[0033] As a further aspect of the present invention: the step of using the adjusted supplementary frame image as the keyframe image, statistically analyzing all keyframe images, and generating an understanding score based on the keyframe images identified by AI includes:
[0034] Use the adjusted supplementary frame image as the keyframe image;
[0035] Analyze all keyframe images to obtain an image sequence;
[0036] Based on AI, image sequences are identified to obtain the recognized text.
[0037] The identified text is compared with the animation script, and the text similarity is calculated as the comprehension score.
[0038] As a further aspect of the present invention: the steps of obtaining the adjustment range of the adjustment information, determining the number of cyclic supplements based on the adjustment range, and cyclically executing the creation process of supplementary frame images until the comprehension reaches a preset comprehension threshold include:
[0039] Obtain the adjustment range input by the designer for the supplementary frame image; the adjustment range is determined by the scaling ratio of all components and the adjustment distance of all components.
[0040] The cyclic replenishment quantity is determined based on the adjustment range; the cyclic replenishment quantity is directly proportional to the adjustment range.
[0041] The number of cyclic supplementation images is used as the number of supplementation frames in the next supplementation frame image construction process;
[0042] The process is repeated until the comprehension level reaches the preset comprehension threshold.
[0043] The present invention also provides a stop-motion animation generation system, the system comprising:
[0044] The text recognition module is used to receive the animation script uploaded by the designer, and sequentially perform text recognition on the sentences in the animation script to obtain the noun groups of each sentence;
[0045] The key statement identification module is used to identify key statements by statistically analyzing noun groups based on the order of statements and analyzing the noun groups.
[0046] The keyframe generation module is used to read noun groups of key sentences, read animation components based on noun groups, and generate keyframe images based on animation components;
[0047] The supplementary frame determination module is used to preserve the statement order, count keyframe images, select adjacent keyframe images in sequence, create and display supplementary frame images based on adjacent keyframe images, and receive adjustment information input by the designer.
[0048] The AI recognition module is used to take the adjusted supplementary frame image as the key frame image, count all key frame images, and generate the comprehension level based on the key frame images counted by AI recognition.
[0049] The loop execution module is used to obtain the adjustment range of the adjustment information, determine the number of loop supplements based on the adjustment range, and loop the creation process of supplementary frame images until the comprehension reaches a preset comprehension threshold.
[0050] As a further aspect of the present invention: the key statement determination module includes:
[0051] Sequential numbering units are used to sequentially number statements and read the noun groups of each statement in turn.
[0052] The word vector generation unit is used to input each noun in the noun group into the word vector conversion model to obtain the word vector of each noun;
[0053] The vector group generation unit is used to count the word vector of each noun and obtain the vector group corresponding to the noun group;
[0054] The feature value generation unit is used to input the vector group into a preset numerical transformation model to obtain the feature values of the vector group;
[0055] The function fitting unit is used to construct coordinate points by taking the number as the independent variable and the feature value as the dependent variable, and fitting the feature value change function.
[0056] The tag extraction unit is used to obtain the derivative of the feature value change function, tag the function segment whose derivative reaches a preset threshold, query the number of the tagged function segment, and use the statement corresponding to the number as the key statement.
[0057] As a further aspect of the present invention: the keyframe generation module includes:
[0058] The noun phrase reading unit is used to read noun phrases from key sentences;
[0059] The component matching unit is used to match components from a preset component library based on each noun, and to use them as animation components;
[0060] The query combination unit is used to query the type of animation components, combine animation components based on the type, and generate keyframe images;
[0061] The display adjustment unit is used to display keyframe images and receive adjustment information input by the designer.
[0062] The present invention also provides a storage medium storing at least one line of program code, which, when loaded and executed by a processor, implements the stop-motion animation generation method.
[0063] Compared with the prior art, the beneficial effects of the present invention are:
[0064] This invention obtains components from the animation script, combines the components to generate keyframe images, and continuously adds supplementary frame images based on the keyframe images. With the help of AI, the existing keyframe images are continuously identified until the comprehensibility reaches the preset conditions. In this process, the designer only needs to upload the script and some fine-tuning instructions for the keyframe images to obtain stop-motion animation. Under the premise that the accuracy requirements are not high, the efficiency is extremely high. Attached Figure Description
[0065] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention.
[0066] Figure 1 The overall flowchart of the method for generating stop-motion animation is shown.
[0067] Figure 2 A structural diagram of the stop-motion animation generation system is shown. Detailed Implementation
[0068] To make the technical problems to be solved, the technical solutions, and the beneficial effects of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.
[0069] Figure 1 This is a flowchart illustrating the method, system, and storage medium for generating stop-motion animation. In this embodiment of the invention, a method for generating stop-motion animation includes:
[0070] Step S100: Receive the animation script uploaded by the designer, and perform text recognition on the sentences in the animation script in turn to obtain the noun group of each sentence;
[0071] The designer is the main body of the stop-motion animation design. The designer uploads the animation script, which contains multiple scenes. Each scene includes multiple paragraphs, and each paragraph includes multiple sentences. In the technical solution of this invention, the smallest unit of the animation script is directly set as a sentence. After receiving the animation script uploaded by the designer, the sentences in the animation script are sequentially subjected to text recognition, and the nouns in the sentences are extracted to obtain the noun phrases corresponding to each sentence, which are called the noun phrases of each sentence.
[0072] Step S200: Based on the sentence order, count the noun groups, identify the noun groups, and determine the key sentences;
[0073] By statistically analyzing the order of sentences in the animation script (sentence order), and comparing the noun groups of adjacent sentences, we can determine the changes in the noun groups, and then select the more critical sentences, which are called key sentences.
[0074] Step S300: Read the noun groups of key statements, read the animation components based on the noun groups, and generate keyframe images based on the animation components;
[0075] The system reads each noun from the noun group of key statements, queries the preset animation component library based on the nouns, counts the obtained animation components, and generates an image called a keyframe image. The animation component library includes noun items and animation component items, which is essentially a component dictionary. In addition, for the same noun, there may be multiple components. In this case, the frequency of use of each component for each noun can be determined based on historical usage records. The propagation degree is determined according to the direct proportion of the frequency of use. The propagation degree requirements input by the designer are received, and then the animation components are selected.
[0076] Step S400: Preserve the statement order, count the keyframe images, select adjacent keyframe images in sequence, create and display supplementary frame images based on the adjacent keyframe images, and receive adjustment information input by the designer;
[0077] Arranging keyframe images based on the order of statements in the animation script yields a stop-motion animation. The more keyframe images, the more coherent and comprehensible the stop-motion animation. Supplementary frame images are determined based on the obtained keyframe images. The method of determination is to automatically add a supplementary frame image between two adjacent keyframe images. It is an intermediate image simulated based on the two preceding and following keyframe images, which may have some differences. Therefore, it is necessary to display the supplementary frame image and then receive adjustment information input by the designer to obtain supplementary frame images that meet the conditions.
[0078] Step S500: Use the adjusted supplementary frame image as the keyframe image, count all keyframe images, and generate the understanding score based on the keyframe images identified by AI.
[0079] The adjusted supplementary frame image is accurate enough to serve as a new keyframe image. Then, supplementary frame images are created based on the new keyframe images. By continuously executing steps S400 and S500, new keyframe images can be continuously generated.
[0080] Step S600: Obtain the adjustment range of the adjustment information, determine the number of cyclic supplements based on the adjustment range, and repeatedly execute the process of creating supplementary frame images until the comprehension reaches a preset comprehension threshold;
[0081] During the continuous generation of new keyframe images, the designer's adjustments to the supplementary frame images are recorded in real time. That is, the amount of adjustment operation made by the designer on the supplementary frame images. Based on the amount of adjustment operation, the accuracy of fitting a new keyframe image based on the current keyframe image can be determined. Generally, the more existing keyframe images there are, the higher the fitting accuracy, and the fewer new supplementary frames need to be created. The cyclic supplementation number mentioned above refers to the number of supplementary frames to be fitted in two adjacent keyframe images. The cyclic supplementation number is directly proportional to the adjustment range; the larger the adjustment range, the larger the cyclic supplementation number.
[0082] In addition, there is a breakout condition in the loop execution process. Each time a new keyframe image is generated, the AI is used to identify all keyframe images to determine the comprehensibility of the current stop-motion animation. When the comprehension is high enough, the loop is broken and all the final keyframe images are output. As a stop-motion animation, this process only requires the designer to upload the animation script and some fine-tuning information, which is extremely convenient and suitable for generating a draft or introduction. If there are detailed requirements, adjustments can be made independently. The independent adjustment process is not part of the technical solution of this invention, so it will not be described in detail.
[0083] Regarding step S100, the animation script uploaded by the designer is received, and the sentences in the animation script are sequentially subjected to text recognition to obtain the noun groups of each sentence. This process can be carried out using existing part-of-speech recognition algorithms to extract nouns from each sentence.
[0084] Regarding step S200, the step of identifying key sentences by statistically analyzing noun groups based on sentence order includes:
[0085] Number the statements sequentially and read the noun phrases in each statement in turn;
[0086] Each noun in the noun group is input into a word vector conversion model to obtain a word vector for each noun;
[0087] Calculate the word vector for each noun to obtain the vector group corresponding to the noun group;
[0088] Input the vector group into the preset numerical transformation model to obtain the feature values of the vector group;
[0089] Using the number as the independent variable and the eigenvalue as the dependent variable, coordinate points are constructed, and the eigenvalue variation function is fitted.
[0090] Obtain the derivative of the eigenvalue change function, mark the function segment whose derivative reaches a preset threshold, query the number of the marked function segment, and use the statement corresponding to the number as the key statement.
[0091] Each statement in the animation script is sequentially numbered. The noun groups of each statement are read one by one. Since noun groups are string sets, processing them is relatively difficult. To simplify the operation, each noun in the noun group is input into a word vector conversion model to obtain a word vector for each noun. The word vectors of each noun are counted to obtain a vector group corresponding to the noun group. At this point, each statement corresponds to a vector group. The purpose of the above is to obtain the changes in nouns in different statements. The vector groups are input into a preset numerical conversion model to obtain the feature values of the vector groups. At this point, each statement corresponds to a numerical value, called a feature value. Using the number as the independent variable and the feature value as the dependent variable, coordinate points are constructed, and a feature value change function is fitted. The derivative function of the feature value change function is obtained. The derivative function reflects the feature value changes of each number (corresponding statement). Function segments where the derivative function reaches a preset threshold are marked. The number of the marked function segment is queried, and the statement corresponding to that number is designated as the key statement.
[0092] Regarding step S300, the steps of reading the noun groups of key statements, reading animation components based on the noun groups, and generating keyframe images based on the animation components include:
[0093] Read the noun phrases in the key sentences;
[0094] For each noun, a component is matched from a preset component library and used as an animation component;
[0095] Query the type of animation component, combine animation components based on type, and generate keyframe images;
[0096] Displays keyframe images and receives adjustment information input by the designer.
[0097] Once the key statements are determined, the noun groups of the key statements are read. For each noun, a component is matched in a preset component library as an animation component. The type of the animation component is queried, and the animation components are combined based on the type to generate keyframe images. The type represents the placement area and display method of the animation component. For example, if the type is "background," it can be displayed at the bottom layer; if it is "land animal," it can be displayed at the top layer and connected to the ground component. The placement area and display method of different types are preset parameters that can be directly read. Even the initially generated keyframe images may contain errors. Therefore, it is necessary to display the keyframe images and receive adjustment information input by the designer. Only after receiving the adjustment information is the keyframe image considered true. Similarly, supplementary frame images can only become keyframe images after receiving adjustment information.
[0098] Regarding step S400, the steps of preserving the statement order, statistically analyzing keyframe images, sequentially selecting adjacent keyframe images, creating and displaying supplementary frame images based on adjacent keyframe images, and receiving adjustment information input by the designer include:
[0099] Preserve statement order and analyze keyframe images;
[0100] Select adjacent keyframe images sequentially, compare the animation components of adjacent keyframe images, and determine the change vector of each animation component;
[0101] The supplementary positions of each animation component are determined based on the change vector, the supplementary positions of all animation components are counted, and a supplementary frame image is constructed.
[0102] Displays supplementary frame images and receives adjustment information input by the designer.
[0103] Specifically, the process of generating supplementary frame images is not complicated. It preserves the order of statements, counts keyframe images, selects adjacent keyframe images in sequence, compares the animation components of adjacent keyframe images, determines the change vector of each animation component, determines the supplementary position of each animation component based on the change vector, generally using the midpoint position of the change vector, counts the supplementary positions of all animation components, constructs supplementary frame images, displays supplementary frame images, and receives adjustment information input by the designer.
[0104] Furthermore, the process of determining the change vector for each animation component includes:
[0105] For animation components that appear in the previous keyframe image but not in the subsequent keyframe image, generate a change vector pointing from the animation component to outside the image frame;
[0106] For an animated component that appears in the previous keyframe image and in the next keyframe image, generate a change vector from the previous position to the next position of the animated component.
[0107] For an animated component that does not appear in the previous keyframe but appears in the next keyframe, generate a change vector pointing from outside the frame to the position of the animated component.
[0108] The above describes three scenarios. First, a component appears only in the previous keyframe and disappears in the next. In this case, the change vector points from the animation component to outside the image frame. If the middle position of the vector is chosen as the replacement position, the distance between the center of the change vector and the image frame must be large enough to reach a preset distance threshold to ensure that the animation component does not obscure the image frame when it is in the center. Second, a component appears only in the next keyframe and does not exist in the previous keyframe. In this case, the change vector points from outside the image frame to the animation component. Similarly, if the middle position of the vector is chosen as the replacement position, the distance between the center of the change vector and the image frame must be large enough to reach a preset distance threshold to ensure that the animation component does not obscure the image frame when it is in the center. Third, the animation component exists in both keyframes. In this case, the change vector is generated from the previous position to the next position, and the center position is selected as the replacement position.
[0109] It's worth mentioning that the first two cases mentioned above refer to the deletion and insertion of components. This method of adjusting components is very common in image processing software such as PowerPoint. As for the third case, it involves selecting the center position as the new position of the animation component. If it is a still life, the center position remains the original position. If it is an animal, the center position is the middle position between the two positions.
[0110] Regarding step S500, the step of using the adjusted supplementary frame image as a keyframe image, counting all keyframe images, and generating an understanding score based on the keyframe images identified by AI includes:
[0111] Use the adjusted supplementary frame image as the keyframe image;
[0112] Analyze all keyframe images to obtain an image sequence;
[0113] Based on AI, image sequences are identified to obtain the recognized text.
[0114] The identified text is compared with the animation script, and the text similarity is calculated as the comprehension score.
[0115] The supplementary frame image after receiving the adjustment information is regarded as a new keyframe image. All keyframe images are counted to obtain an image sequence. Based on the AI recognition of the image sequence, the recognized text is obtained. The recognized text is compared with the animation script, and the text similarity is calculated as the comprehension level. Under the current comprehension capabilities of AI, this process is very simple. The text similarity calculation process can be carried out using existing text comparison algorithms.
[0116] Regarding step S600, the steps of obtaining the adjustment range of the adjustment information, determining the number of cyclic supplements based on the adjustment range, and cyclically executing the process of creating supplementary frame images until the comprehension level reaches a preset comprehension threshold include:
[0117] Obtain the adjustment range input by the designer for the supplementary frame image; the adjustment range is determined by the scaling ratio of all components and the adjustment distance of all components.
[0118] The cyclic replenishment quantity is determined based on the adjustment range; the cyclic replenishment quantity is directly proportional to the adjustment range.
[0119] The number of cyclic supplementation images is used as the number of supplementation frames in the next supplementation frame image construction process;
[0120] The process is repeated until the comprehension level reaches the preset comprehension threshold.
[0121] The main application of the technical solution of this invention is generally a computer device. After generating the supplementary frame image, the computer device needs to receive the adjustment information input by the designer. After receiving the adjustment information, the supplementary frame image is used as the new keyframe image. At the same time, the adjustment range of the adjustment information is identified, and the number of cyclic supplements is determined according to the adjustment range. The number of cyclic supplements is used as the number of supplementary frame images in the next supplementary frame image construction process. For example, when the number of cyclic supplements is two, two supplementary frame images need to be inserted into adjacent keyframe images. At this time, it is necessary to select one-third of the position on the change vector as the supplementary position.
[0122] The process of generating supplementary frame images is repeated until the comprehension level reaches a preset comprehension threshold.
[0123] Figure 2 A structural diagram of a stop-motion animation generation system is shown. In a preferred embodiment of the technical solution of the present invention, a stop-motion animation generation system is also provided, the system 10 comprising:
[0124] The text recognition module 11 is used to receive the animation script uploaded by the designer, and to perform text recognition on the sentences in the animation script in turn to obtain the noun group of each sentence.
[0125] The key statement identification module 12 is used to identify key statements by statistically analyzing noun groups based on the statement order and identifying the noun groups.
[0126] The keyframe generation module 13 is used to read noun groups of key sentences, read animation components based on noun groups, and generate keyframe images based on animation components;
[0127] The supplementary frame determination module 14 is used to preserve the statement order, count key frame images, select adjacent key frame images in sequence, create and display supplementary frame images based on adjacent key frame images, and receive adjustment information input by the designer.
[0128] AI recognition module 15 is used to take the adjusted supplementary frame image as the key frame image, count all key frame images, and generate comprehension based on the key frame images counted by AI recognition.
[0129] The loop execution module 16 is used to obtain the adjustment range of the adjustment information, determine the number of loop supplements based on the adjustment range, and loop the creation process of the supplementary frame image until the comprehension reaches a preset comprehension threshold.
[0130] Furthermore, the key statement determination module 12 includes:
[0131] Sequential numbering units are used to sequentially number statements and read the noun groups of each statement in turn.
[0132] The word vector generation unit is used to input each noun in the noun group into the word vector conversion model to obtain the word vector of each noun;
[0133] The vector group generation unit is used to count the word vector of each noun and obtain the vector group corresponding to the noun group;
[0134] The feature value generation unit is used to input the vector group into a preset numerical transformation model to obtain the feature values of the vector group;
[0135] The function fitting unit is used to construct coordinate points by taking the number as the independent variable and the feature value as the dependent variable, and fitting the feature value change function.
[0136] The tag extraction unit is used to obtain the derivative of the feature value change function, tag the function segment whose derivative reaches a preset threshold, query the number of the tagged function segment, and use the statement corresponding to the number as the key statement.
[0137] Specifically, the keyframe generation module 13 includes:
[0138] The noun phrase reading unit is used to read noun phrases from key sentences;
[0139] The component matching unit is used to match components from a preset component library based on each noun, and to use them as animation components;
[0140] The query combination unit is used to query the type of animation components, combine animation components based on the type, and generate keyframe images;
[0141] The display adjustment unit is used to display keyframe images and receive adjustment information input by the designer.
[0142] The above are merely preferred embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.
Claims
1. A method for generating stop-motion animation, characterized in that, The method includes: The system receives animation scripts uploaded by designers, performs text recognition on the sentences in the animation scripts in turn, and obtains the noun groups of each sentence; Based on the sentence order, noun groups are statistically analyzed to identify key sentences. Read the noun phrases of key statements, read the animation components based on the noun phrases, and generate keyframe images based on the animation components; Preserve the statement order, count keyframe images, select adjacent keyframe images in sequence, create and display supplementary frame images based on adjacent keyframe images, and receive adjustment information input by the designer; The adjusted supplementary frame images are used as keyframe images. All keyframe images are counted, and the understanding is generated based on the keyframe images identified by AI. The adjustment range of the adjustment information is obtained, the number of cyclic supplements is determined according to the adjustment range, and the process of creating supplementary frame images is executed cyclically until the comprehension reaches a preset comprehension threshold. During the continuous generation of new keyframe images, the amount of adjustment operations made by the designer on the supplementary frame images is recorded in real time. The accuracy of fitting a new keyframe image based on the current keyframe image is determined based on the amount of adjustment operations. The number of cyclic supplements is the number of supplementary frames to be fitted in two adjacent keyframe images. The number of cyclic supplements is directly proportional to the adjustment range. The larger the adjustment range, the larger the number of cyclic supplements. The loop execution process is terminated as follows: each time a new keyframe image is generated, all keyframe images are identified based on AI to determine the comprehensibility of the current stop-motion animation. When the comprehensibility reaches the preset comprehensibility threshold, the loop is terminated and all keyframe images are output as the final stop-motion animation.
2. The method for generating stop-motion animation according to claim 1, characterized in that, The steps for identifying key sentences by statistically analyzing noun groups based on sentence order include: Number the statements sequentially and read the noun phrases in each statement in turn; Each noun in the noun group is input into a word vector conversion model to obtain a word vector for each noun; Calculate the word vector for each noun to obtain the vector group corresponding to the noun group; Input the vector group into the preset numerical transformation model to obtain the feature values of the vector group; Using the number as the independent variable and the eigenvalue as the dependent variable, coordinate points are constructed, and the eigenvalue variation function is fitted. Obtain the derivative of the eigenvalue change function, mark the function segment whose derivative reaches a preset threshold, query the number of the marked function segment, and use the statement corresponding to the number as the key statement.
3. The method for generating stop-motion animation according to claim 1, characterized in that, The steps of reading noun groups from key statements, reading animation components based on noun groups, and generating keyframe images based on animation components include: Read the noun phrases in the key sentences; For each noun, a component is matched from a preset component library and used as an animation component; Query the type of animation component, combine animation components based on type, and generate keyframe images; Displays keyframe images and receives adjustment information input by the designer.
4. The method for generating stop-motion animation according to claim 1, characterized in that, The steps of preserving the statement order, counting keyframe images, sequentially selecting adjacent keyframe images, creating and displaying supplementary frame images based on adjacent keyframe images, and receiving adjustment information input by the designer include: Preserve statement order and analyze keyframe images; Select adjacent keyframe images sequentially, compare the animation components of adjacent keyframe images, and determine the change vector of each animation component; The supplementary positions of each animation component are determined based on the change vector, the supplementary positions of all animation components are counted, and a supplementary frame image is constructed. Displays supplementary frame images and receives adjustment information input by the designer; The process of determining the change vector for each animation component includes: For animation components that appear in the previous keyframe image but not in the subsequent keyframe image, generate a change vector pointing from the animation component to outside the image frame; For an animated component that appears in the previous keyframe image and in the next keyframe image, generate a change vector from the previous position to the next position of the animated component. For an animated component that does not appear in the previous keyframe but appears in the next keyframe, generate a change vector pointing from outside the frame to the position of the animated component.
5. The method for generating stop-motion animation according to claim 1, characterized in that, The steps of using the adjusted supplementary frame image as the keyframe image, counting all keyframe images, and generating the understanding score based on the keyframe images identified by AI include: Use the adjusted supplementary frame image as the keyframe image; Analyze all keyframe images to obtain an image sequence; Based on AI, image sequences are identified to obtain the recognized text. The identified text is compared with the animation script, and the text similarity is calculated as the comprehension score.
6. The method for generating stop-motion animation according to claim 1, characterized in that, The steps of obtaining the adjustment range of the adjustment information, determining the number of cyclic supplements based on the adjustment range, and cyclically executing the process of creating supplementary frame images until the comprehension level reaches a preset comprehension threshold include: Obtain the adjustment range input by the designer for the supplementary frame image; the adjustment range is determined by the scaling ratio of all components and the adjustment distance of all components. The cyclic replenishment quantity is determined based on the adjustment range; the cyclic replenishment quantity is directly proportional to the adjustment range. The number of cyclic supplementation images is used as the number of supplementation frames in the next supplementation frame image construction process; The process is repeated until the comprehension level reaches the preset comprehension threshold.
7. A stop-motion animation generation system, characterized in that, The system includes: The text recognition module is used to receive the animation script uploaded by the designer, and sequentially perform text recognition on the sentences in the animation script to obtain the noun groups of each sentence; The key statement identification module is used to identify key statements by statistically analyzing noun groups based on the order of statements and analyzing the noun groups. The keyframe generation module is used to read noun groups of key sentences, read animation components based on noun groups, and generate keyframe images based on animation components; The supplementary frame determination module is used to preserve the statement order, count keyframe images, select adjacent keyframe images in sequence, create and display supplementary frame images based on adjacent keyframe images, and receive adjustment information input by the designer. The AI recognition module is used to take the adjusted supplementary frame image as the key frame image, count all key frame images, and generate the comprehension level based on the key frame images counted by AI recognition. The loop execution module is used to obtain the adjustment range of the adjustment information, determine the number of loop supplements based on the adjustment range, and loop the creation process of the supplementary frame image until the comprehension reaches a preset comprehension threshold. During the continuous generation of new keyframe images, the amount of adjustment operations made by the designer on the supplementary frame images is recorded in real time. The accuracy of fitting a new keyframe image based on the current keyframe image is determined based on the amount of adjustment operations. The number of cyclic supplements is the number of supplementary frames to be fitted in two adjacent keyframe images. The number of cyclic supplements is directly proportional to the adjustment range. The larger the adjustment range, the larger the number of cyclic supplements. The loop execution process is terminated as follows: each time a new keyframe image is generated, all keyframe images are identified based on AI to determine the comprehensibility of the current stop-motion animation. When the comprehensibility reaches the preset comprehensibility threshold, the loop is terminated and all keyframe images are output as the final stop-motion animation.
8. The stop-motion animation generation system according to claim 7, characterized in that, The key statement determination module includes: Sequential numbering units are used to sequentially number statements and read the noun groups of each statement in turn. The word vector generation unit is used to input each noun in the noun group into the word vector conversion model to obtain the word vector of each noun; The vector group generation unit is used to count the word vector of each noun and obtain the vector group corresponding to the noun group; The feature value generation unit is used to input the vector group into a preset numerical transformation model to obtain the feature values of the vector group; The function fitting unit is used to construct coordinate points by taking the number as the independent variable and the feature value as the dependent variable, and fitting the feature value change function. The tag extraction unit is used to obtain the derivative of the feature value change function, tag the function segment whose derivative reaches a preset threshold, query the number of the tagged function segment, and use the statement corresponding to the number as the key statement.
9. The stop-motion animation generation system according to claim 7, characterized in that, The keyframe generation module includes: The noun phrase reading unit is used to read noun phrases from key sentences; The component matching unit is used to match components from a preset component library based on each noun, and to use them as animation components; The query combination unit is used to query the type of animation components, combine animation components based on the type, and generate keyframe images; The display adjustment unit is used to display keyframe images and receive adjustment information input by the designer.
10. A storage medium, characterized in that, The storage medium stores at least one piece of program code, which, when loaded and executed by a processor, implements the stop-motion animation generation method as described in any one of claims 1-6.