Cartoon video medium-based poster batch generation method and system

Abstract

The invention discloses a poster batch generation method and system based on a cartoon video medium, and the method comprises the steps: obtaining the configuration parameters of a one-card video, and the configuration parameters comprise role information and text information; performing image conversion processing on the cartoon video to obtain an image set; identifying the role information of each image in the image set by using a preset model to obtain a role image set; acquiring a poster image set of the cartoon video by utilizing a role position in the role image set and an adding position of the text information; and outputting the poster image set. According to the poster batch generation method and system based on the cartoon video medium, the poster pictures of the cartoon video can be automatically generated, the forms of the poster pictures are various, the production efficiency is greatly improved, the method and system can automatically adapt to the identification process of animation images, and the success rate of the poster is improved.

Description

Technical Field

[0001] This invention relates to a method and system for batch generation of posters based on cartoon video media. Background Technology

[0002] Upstream video platforms refer to those focused on professional content production and copyright integration. They invest in self-production or acquisition of film, television, and variety show content to create content libraries, acting as the "content source" and "copyright holder" of the video industry chain. Downstream video platforms, on the other hand, primarily rely on user-generated content and algorithm-based distribution and display. Their core function is building a creative ecosystem and recommendation system, aggregating traffic by stimulating user creation and accurately pushing content, playing the roles of "content distribution amplifier" and "social interaction arena." Together, they constitute the modern video ecosystem, and their boundaries are increasingly merging as short video platforms expand into long-form video and traditional platforms introduce short video features.

[0003] In existing technologies, downstream video platforms require multiple episode posters, but upstream copyright holders do not provide corresponding posters. This is generally handled manually, which is inefficient. This is especially true for cartoon videos, which are often divided into seasons with a large number of episodes per season. Therefore, poster generation for cartoon videos is typically done manually, resulting in low efficiency and limited design options. Summary of the Invention

[0004] The technical problem to be solved by this invention is to overcome the shortcomings of existing technologies where posters based on cartoon video media are generated manually, resulting in low generation efficiency and limited styles. This invention provides a method and system for batch generation of posters based on cartoon video media that can automatically generate poster images of cartoon videos, with diverse poster image formats, significantly improving production efficiency, automatically adapting to the recognition process of animated images, and increasing the success rate of posters.

[0005] This application solves the above-mentioned technical problems through the following technical solution:

[0006] A method for batch generating posters based on cartoon video media, characterized in that the method includes:

[0007] Obtain the configuration parameters of the all-in-one card video, the configuration parameters including role information and text information;

[0008] The cartoon video is subjected to image conversion processing to obtain an image set;

[0009] A character image set is obtained by recognizing the character information in each image in the image set using a preset model.

[0010] The poster image set of the cartoon video is obtained by utilizing the character positions in the character image set and the positions where text information is added;

[0011] Output the set of poster images.

[0012] Preferably, the step of using a preset model to identify the character information in each image in the image set to obtain the character image set includes, prior to:

[0013] Obtain the training image set for the character information;

[0014] The preset model is trained using the training image set.

[0015] Preferably, the method for batch generating posters includes:

[0016] Collect user editing data, which includes the number of characters, character information, text information, and poster format information;

[0017] The configuration parameters of the cartoon video are obtained using the edited data.

[0018] Preferably, the editing data includes the number of images in the poster image set, and the poster batch generation method includes:

[0019] The image set is frame-by-frame extracted at a first preset interval to obtain a first image subset;

[0020] The first character image set is obtained by recognizing the character information in each image in the first image subset using a preset model, and the recognition time parameter is obtained, which includes the average recognition time and / or the number of images with a recognition time greater than the preset recognition time.

[0021] The duration of the second preset interval is obtained based on the recognition duration parameter and the number of images. The duration of the second preset interval is shorter than that of the first preset interval. The larger the recognition duration parameter, the shorter the duration of the second preset interval. The more difficult the image to recognize, the more complex the information content, the less suitable it is for making a poster, and the more image sets are needed.

[0022] The image set is frame-by-frame extracted at a second preset interval to obtain a second image subset;

[0023] Combine the first image subset and the second image subset to obtain the image set.

[0024] Preferably, the method for batch generating posters includes:

[0025] The images with captions in the image set were identified using OCR technology.

[0026] Remove images with captions to get the latest image set;

[0027] The character image set is obtained by recognizing the character information in each image in the latest image set using a preset model.

[0028] Preferably, the method for batch generating posters includes:

[0029] The images in the image set are processed into black and white;

[0030] The image set containing black and white subtitles was identified using OCR technology.

[0031] Delete the image with captions;

[0032] The closest blank caption image before and after the deleted image in the cartoon video is obtained and added to the image set to obtain the latest image set.

[0033] Preferably, the step of obtaining the poster image set of the cartoon video by utilizing the character positions in the character image set and the addition positions of text information includes:

[0034] For each character image in the character image set, obtain the absolute distance from the character outline to the image boundary;

[0035] Delete character images whose absolute distance is less than a preset value to obtain the latest character image set;

[0036] The poster image set of the cartoon video is obtained by utilizing the character positions and text information addition positions in the latest character image set.

[0037] Preferably, obtaining the poster image set of the cartoon video using the character positions and text information addition positions in the latest character image set includes:

[0038] The size and position of the text information are set according to the character's position;

[0039] Add text information indicating the target size to the target location in the character image;

[0040] The character image with added text information is cropped according to the position of the character in the character image and the positional relationship of the text information to obtain the poster image corresponding to the character image.

[0041] Preferably, FFmpeg is called to perform image conversion processing on the cartoon video, and the YOLO model is used to identify the character information in each image in the image set to obtain a character image set.

[0042] A poster batch generation system based on cartoon video media is characterized in that the poster batch generation system is used to implement the poster batch generation method based on cartoon video media as described above.

[0043] Based on common knowledge in the field, the above-mentioned preferred conditions can be combined arbitrarily to obtain various preferred embodiments of the present invention.

[0044] The positive and progressive effects of this invention are as follows:

[0045] The poster batch generation method and system based on cartoon video media of the present invention can automatically generate poster images of cartoon videos, and the poster images are diverse in form, which greatly improves production efficiency and can automatically adapt to the recognition process of animation images, thereby improving the success rate of posters. Attached Figure Description

[0046] Figure 1 This is a flowchart of the poster batch generation method according to Embodiment 1 of the present invention. Detailed Implementation

[0047] The present invention will be further illustrated by way of embodiments below, but the present invention is not limited to the scope of the embodiments described herein.

[0048] Example 1

[0049] This embodiment provides a poster batch generation system based on cartoon video media. The poster batch generation system includes a mobile terminal and a server. The mobile terminal can be a mobile phone or a tablet computer.

[0050] The mobile terminal is used to obtain the configuration parameters of the all-in-one card video, and the configuration parameters include role information and text information;

[0051] The server is used for:

[0052] Receive the configuration parameters;

[0053] The cartoon video is subjected to image conversion processing to obtain an image set;

[0054] A character image set is obtained by recognizing the character information in each image in the image set using a preset model.

[0055] The poster image set of the cartoon video is obtained by utilizing the character positions in the character image set and the positions where text information is added;

[0056] The poster image set is output to the mobile terminal.

[0057] The server is used for:

[0058] Obtain the training image set of the role information sent by the mobile terminal;

[0059] The preset model is trained using the training image set.

[0060] The mobile terminal is used for:

[0061] Collect user editing data, which includes the number of characters, character information, text information, and poster format information;

[0062] The configuration parameters of the cartoon video are obtained using the edited data.

[0063] The server is used for:

[0064] The image set is frame-by-frame extracted at a first preset interval to obtain a first image subset;

[0065] The first character image set is obtained by recognizing the character information in each image in the first image subset using a preset model, and the recognition time parameter is obtained, which includes the average recognition time and / or the number of images with a recognition time greater than the preset recognition time.

[0066] The duration of the second preset interval is obtained based on the recognition duration parameter and the number of images. The duration of the second preset interval is shorter than that of the first preset interval. The larger the recognition duration parameter, the shorter the duration of the second preset interval. The more difficult the image to recognize, the more complex the information content, the less suitable it is for making a poster, and the more image sets are needed.

[0067] The image set is frame-by-frame extracted at a second preset interval to obtain a second image subset;

[0068] Combine the first image subset and the second image subset to obtain the image set.

[0069] The server is used for:

[0070] The images with captions in the image set were identified using OCR technology.

[0071] Remove images with captions to get the latest image set;

[0072] The character image set is obtained by recognizing the character information in each image in the latest image set using a preset model.

[0073] The server is used for:

[0074] The images in the image set are processed into black and white;

[0075] The image set containing black and white subtitles was identified using OCR technology.

[0076] Delete the image with captions;

[0077] The closest blank caption image before and after the deleted image in the cartoon video is obtained and added to the image set to obtain the latest image set.

[0078] Among them, blank letter images refer to images without subtitles.

[0079] The server is used for:

[0080] For each character image in the character image set, obtain the absolute distance from the character outline to the image boundary;

[0081] The absolute distance includes the distance from the character's outline to each boundary.

[0082] Delete character images whose absolute distance is less than a preset value to obtain the latest character image set;

[0083] The poster image set of the cartoon video is obtained by utilizing the character positions and text information addition positions in the latest character image set.

[0084] The server is used for:

[0085] The size and position of the text information are set according to the character's position;

[0086] Add text information indicating the target size to the target location in the character image;

[0087] The character image with added text information is cropped according to the position of the character in the character image and the positional relationship of the text information to obtain the poster image corresponding to the character image.

[0088] The cartoon video is processed by calling FFmpeg to perform image conversion, and the character image set is obtained by recognizing the character information in each image in the image set using the YOLO model.

[0089] In this embodiment, the user operation steps are as follows:

[0090] 1. Select the same folder as the previous step;

[0091] 2. Selection rules (characters, occurrence conditions) (e.g., a character must appear, any N characters, etc.);

[0092] 3. Configure specific fonts or images;

[0093] 4. Configure horizontal and vertical cropping rules;

[0094] 5. Configure the number of posters;

[0095] 6. Configure the interval for retrieving images (too small an interval will result in a high degree of image similarity; the interval for retrieving images should be matched with the minimum number of posters).

[0096] 7. Submit the task to the server;

[0097] 8. After the background process is complete, download the automatically generated poster compressed package and manually select the posters.

[0098] The server-side steps are as follows:

[0099] 1. Read the video file and use code to call FFmpeg (a powerful open-source audio and video processing tool) to convert the video into images;

[0100] 2. Remove images with subtitles (after preprocessing the images using OCR technology, perform text recognition and discard images with recognized text). Delete any images with text, process them in black and white, and then perform OCR again.

[0101] 3. Call YOLO and use the trained model to identify people in the image. Based on the location information of the identified people, keep only the selected people that meet the number requirement, and discard the other images.

[0102] 4. Crop (filter) the images according to the cropping settings (horizontal and vertical aspect ratios, i.e., the aspect ratios and settings are different for each site). At the same time, record the relative position of the people after cropping. Discard images that are too off-center after cropping, or images that are not tall enough or short enough (consider that adding a bottom or top label might obscure the people, so these are also unacceptable).

[0103] 5. Scale the image of the artistic text and place it in the configured position. Record the relative position of the placement and compare it with the position of the person's front in step 4. Discard the image that is obscured by the position.

[0104] 6. Once the quantity requirement is met or the video ends, package the generated images together.

[0105] See Figure 1 Using the aforementioned poster batch generation system based on cartoon video media, this embodiment also provides a poster batch generation method, including:

[0106] Step 100: Collect user editing data, including the number of characters, character information, text information, and poster format information;

[0107] Step 101: Use the edited data to obtain the configuration parameters of the cartoon video.

[0108] Step 102: Obtain the configuration parameters of the all-in-one card video, including role information and text information;

[0109] Step 103: Perform image conversion processing on the cartoon video to obtain an image set;

[0110] Step 104: Use a preset model to identify the character information in each image in the image set to obtain a character image set;

[0111] Step 105: Obtain the poster image set of the cartoon video by using the character positions in the character image set and the positions where text information is added;

[0112] Step 106: Output the poster image set.

[0113] Step 104 includes:

[0114] Obtain the training image set for the character information;

[0115] The preset model is trained using the training image set.

[0116] The above steps can be performed simultaneously in step 101.

[0117] The editing data includes the number of images in the poster image set, and the poster batch generation method includes:

[0118] The image set is frame-by-frame extracted at a first preset interval to obtain a first image subset;

[0119] The first character image set is obtained by recognizing the character information in each image in the first image subset using a preset model, and the recognition time parameter is obtained, including the average recognition time.

[0120] The duration of the second preset interval is obtained based on the recognition duration parameter and the number of images. The duration of the second preset interval is less than that of the first preset interval, and the larger the recognition duration parameter is, the shorter the duration of the second preset interval is.

[0121] The recognition time parameter also includes the number of images with a recognition time greater than the preset time, which reflects the recognition time. The more difficult the image is to recognize, the more complex the information content, the longer the time required, the less suitable it is for posters, and the more images are needed.

[0122] The image set is frame-by-frame extracted at a second preset interval to obtain a second image subset;

[0123] Combine the first image subset and the second image subset to obtain the image set, and then execute step 104 again.

[0124] The method for batch generating posters includes:

[0125] The images with captions in the image set were identified using OCR technology.

[0126] Remove images with captions to get the latest image set;

[0127] The character image set is obtained by recognizing the character information in each image in the latest image set using a preset model.

[0128] The method for batch generating posters includes:

[0129] The images in the image set are processed into black and white;

[0130] The image set containing black and white subtitles was identified using OCR technology.

[0131] Delete the image with captions;

[0132] The closest blank caption image before and after the deleted image in the cartoon video is obtained and added to the image set to obtain the latest image set.

[0133] Step 104 includes:

[0134] For each character image in the character image set, obtain the absolute distance from the character outline to the image boundary;

[0135] Delete character images whose absolute distance is less than a preset value to obtain the latest character image set;

[0136] The poster image set of the cartoon video is obtained by utilizing the character positions and text information addition positions in the latest character image set.

[0137] Furthermore, the step of obtaining the poster image set of the cartoon video by utilizing the character positions and text information addition positions in the latest character image set includes:

[0138] The size and position of the text information are set according to the character's position;

[0139] Add text information indicating the target size to the target location in the character image;

[0140] The character image with added text information is cropped according to the position of the character in the character image and the positional relationship of the text information to obtain the poster image corresponding to the character image.

[0141] While specific embodiments of the present invention have been described above, those skilled in the art should understand that these are merely illustrative examples, and the scope of protection of the present invention is defined by the appended claims. Those skilled in the art can make various changes or modifications to these embodiments without departing from the principles and essence of the present invention, but all such changes and modifications fall within the scope of protection of the present invention.

Claims

1. A method for mass production of posters based on cartoon video media, characterized in that, The poster batch generation method comprises: Obtaining configuration parameters of the cartoon video, the configuration parameters comprising role information and text information; Conducting image conversion processing on the cartoon video to obtain an image set; Using a preset model to identify the role information of each image in the image set to obtain a role image set; Using the role position in the role image set and the adding position of the text information to obtain a poster image set of the cartoon video; Outputting the poster image set.

2. The method of claim 1, wherein the cartoon video-based poster batch generation method is characterized by, Before the step of using a preset model to identify the role information of each image in the image set to obtain a role image set, the method comprises: Obtaining a training image set of the role information; Training the preset model using the training image set.

3. The method of claim 1, wherein the cartoon video-based poster batch generation method is characterized by, The poster batch generation method comprises: Collecting user editing data, the editing data comprising the number of roles, role information, text information and poster format information; Obtaining configuration parameters of the cartoon video using the editing data.

4. The method of claim 3, wherein the cartoon video medium-based poster batch generation method is characterized by, The editing data comprises the number of images in the poster image set, and the poster batch generation method comprises: Frame extraction is performed on the image set at a first preset interval to obtain a first image subset; A preset model is used to identify the role information of each image in the first image subset to obtain a first role image set, and an identification time length parameter is obtained, the identification time length parameter comprising an average identification time length and / or the number of identification time lengths greater than a preset identification time length; According to the identification time length parameter and the number of images, the time length of a second preset interval is obtained, the second preset interval being shorter than the first preset interval, and the greater the identification time length parameter, the shorter the time length of the second preset interval; Frame extraction is performed on the image set at the second preset interval to obtain a second image subset; The first image subset and the second image subset are combined to obtain the image set.

5. The method of claim 1, wherein the cartoon video-based poster batch generation method is characterized by, The poster batch generation method comprises: Using OCR technology to identify images with subtitles in the image set; Deleting the images with subtitles to obtain a latest image set; Using a preset model to identify the role information of each image in the latest image set to obtain a role image set.

6. The method of claim 5, wherein the cartoon video medium-based poster batch generation method is characterized by, The poster batch generation method comprises: Conducting black and white processing on the images in the image set; Using OCR technology to identify black and white images with subtitles in the image set; Deleting the images with subtitles; Obtaining a blank subtitle image closest in time to the deleted images in the cartoon video and adding it to the image set to obtain a latest image set.

7. The method of claim 1, wherein the cartoon video-based poster batch generation method is characterized by, The step of using the role position in the role image set and the adding position of the text information to obtain a poster image set of the cartoon video comprises: For the role images in the role image set, obtaining the absolute distance from the role contour to the image boundary; Deleting the role images with an absolute distance less than a preset value to obtain a latest role image set; Using the role position in the latest role image set and the adding position of the text information to obtain a poster image set of the cartoon video.

8. The method of claim 7, wherein the cartoon video-based poster batch generation method is characterized by, The step of using the role position in the latest role image set and the adding position of the text information to obtain a poster image set of the cartoon video comprises: Setting the size and position of the text information according to the role position; Adding text information of a target size at a target position in a role image; According to the position relationship between the role position in the role image and the position of the text information, the role image to which the text information is added is cropped to obtain a poster image corresponding to the role image.

9. The method of claim 1, wherein the cartoon video-based poster batch generation method is characterized by, Calling FFmpeg to perform image conversion processing on the cartoon video, and using a yolo model to identify the role information in each image in the image set to obtain a role image set.

10. A system for mass production of posters based on cartoon video media, characterized by, The poster batch generation system is used to implement the cartoon video medium-based poster batch generation method according to any one of claims 1 to 9.

Images