Picture generation method and device

Abstract

The invention discloses a picture generation method and device, and the method comprises the steps: obtaining picture generation information which comprises building effect picture data and a description text; determining at least one standby picture from the building effect picture data according to the description text; generating a target webpage file according to the standby picture and the description text; and performing screenshot based on the target webpage file to obtain a target picture.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of picture generation, and particularly relates to a picture generation method and device. BACKGROUND

[0002] When a user uses an electronic device, sometimes various picture posters need to be made, and the picture posters generally include promotional text and a plurality of product-related pictures, such as house type pictures, building appearance pictures, and traffic condition pictures.

[0003] At present, some AI models based on artificial intelligence technology can automatically generate the above picture posters according to the needs of the user, but the generated picture posters often have various defects, such as poor text and background fusion, specifically manifested as font blurring and / or position deviation; and multi-picture splicing logic confusion, specifically manifested as a proportioning disorder of an effect picture and a house type picture. SUMMARY

[0004] In order to improve the quality of the automatically generated picture posters, the present application discloses the following technical solutions:

[0005] The first aspect of the present application provides a picture generation method, comprising:

[0006] obtaining picture generation information, wherein the picture generation information includes building effect picture data and description text;

[0007] determining at least one standby picture from the building effect picture data according to the description text;

[0008] generating a target web page file according to the standby picture and the description text;

[0009] performing screenshot based on the target web page file to obtain a target picture.

[0010] Optionally, the performing screenshot based on the target web page file to obtain a target picture comprises:

[0011] performing compliance detection on the text contained in the target web page file;

[0012] in response to detecting non-compliant text that does not meet a preset text specification condition, replacing the non-compliant text with compliant text having the same semantics or similar semantics to obtain a replacement web page file;

[0013] performing screenshot in a display interface displaying the replacement web page file to obtain a target picture.

[0014] Optionally, the performing screenshot based on the target web page file to obtain a target picture comprises:

[0015] obtaining an adjustment instruction;

[0016] adjust a web page layout of the target web page file according to the adjustment instruction, the web page layout comprising at least one of a text size, a text position, a picture size and a picture position, to obtain an adjusted web page file;

[0017] screenshot a display interface in which the adjusted web page file is displayed, to obtain a target picture.

[0018] Optionally, the description text comprises a structure text, a style text and a content text.

[0019] The generating the target web page file according to the backup picture and the description text comprises:

[0020] processing the content text according to the style text to obtain to-be-displayed text having a target style described by the style text and consistent with the content text;

[0021] formatting the to-be-displayed text and the backup picture according to a format structure described by the structure text to obtain the target web page file.

[0022] Optionally, the determining at least one backup picture from the real estate effect picture data according to the description text comprises:

[0023] selecting at least one target initial picture from a plurality of initial pictures contained in the real estate effect picture data according to the description text;

[0024] matching a target picture style from a plurality of preset picture styles according to the description text;

[0025] processing the at least one target initial picture according to the target picture style to obtain at least one backup picture having the target picture style.

[0026] The second aspect of the present application provides a picture generating apparatus, comprising:

[0027] an obtaining unit configured to obtain picture generating information, the picture generating information comprising real estate effect picture data and description text;

[0028] a determining unit configured to determine at least one backup picture from the real estate effect picture data according to the description text;

[0029] a generating unit configured to generate a target web page file according to the backup picture and the description text;

[0030] a screenshot unit configured to perform screenshot based on the target web page file to obtain a target picture.

[0031] Optionally, the screenshot unit performs screenshot based on the target webpage file to obtain the target picture, and specifically for:

[0032] compliance detection is performed on the text contained in the target webpage file;

[0033] In response to detecting non-compliant text that does not meet the preset text specification condition, the non-compliant text is replaced with compliant text having the same semantics or similar semantics to obtain a replacement webpage file;

[0034] screenshot in a display interface displaying the replacement webpage file to obtain the target picture.

[0035] Optionally, the screenshot unit performs screenshot based on the target webpage file to obtain the target picture, and specifically for:

[0036] obtain an adjustment instruction;

[0037] adjust the webpage layout of the target webpage file according to the adjustment instruction to obtain an adjusted webpage file, the webpage layout including at least one of text size, text position, picture size, and picture position;

[0038] screenshot in a display interface displaying the adjusted webpage file to obtain the target picture.

[0039] Optionally, the description text includes structural text, style text, and content text;

[0040] The generation unit generates a target webpage file according to the backup picture and the description text, and specifically for:

[0041] processing the content text according to the style text to obtain display text having a target style described by the style text and consistent content with the content text;

[0042] layout the display text and the backup picture according to the layout structure described by the structural text to obtain the target webpage file.

[0043] Optionally, the determination unit determines at least one backup picture from the real estate effect picture data according to the description text, and specifically for:

[0044] select at least one target initial picture from a plurality of initial pictures contained in the real estate effect picture data according to the description text;

[0045] match a target picture style from a plurality of preset picture styles according to the description text;

[0046] According to the target picture style processing the at least one target initial picture, at least one backup picture with the target picture style is obtained. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description only constitute the embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of the provided drawings.

[0048] Figure 1 is a flow chart of a picture generation method provided by the embodiments of the present application;

[0049] Figure 2 is a method flow chart for determining a backup picture provided by the embodiments of the present application;

[0050] Figure 3 is a flow chart of a target web page file generation method provided by the embodiments of the present application;

[0051] Figure 4 is a flow chart of a target picture obtained by screenshot provided by the embodiments of the present application;

[0052] Figure 5 is another flow chart of a target picture obtained by screenshot provided by the embodiments of the present application;

[0053] Figure 6 is a structural schematic diagram of a picture generation device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments only constitute some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0055] The embodiments of the present application provide a picture generation method, please refer to Figure 1 , the method can include the following steps.

[0056] S101, obtain picture generation information, the picture generation information includes building effect picture data and description text.

[0057] S102, determining at least one backup picture from the building effect picture data according to the description text.

[0058] S103, generating a target webpage file according to the backup picture and the description text.

[0059] S104, performing screenshot based on the target webpage file to obtain a target picture.

[0060] The embodiment has the beneficial effect that, compared with directly generating a picture by using an AI model, the scheme can better fuse the backup picture and the description text through a webpage file, avoid problems such as font blurring and / or position offset, and effect picture and house type picture proportion imbalance, and improve the quality of the generated picture.

[0061] In S101, the building effect picture data can include a building effect plan previously stored in a knowledge base, a building three-dimensional model, and other images and / or video data that can be related to the specified building, without limitation. Among them, the building effect plan and other images and / or video data that can be related to the specified building are collectively referred to as initial pictures contained in the building effect picture data.

[0062] The embodiment can be executed by any computer system (hereinafter referred to as system), and the system can select the specified building effect picture data in the pre-stored data based on the selection operation of the user, or can search for the specified building effect picture data based on the description text input by the user for the building.

[0063] Optionally, the description text can include any one or more of structure text, style text, and content text.

[0064] The structure text is used to describe the layout of the picture and the text in the generated target webpage file, for example, the text can be displayed above the picture, the text can be embedded in the center of the picture, and multiple pictures can be displayed side by side.

[0065] The style text is used to describe the style of the generated target webpage file, for example, it can be modern minimalist style, retro nostalgic style, etc.

[0066] The content text refers to the text that needs to be displayed in the webpage when generating the target webpage file.

[0067] The description text can be directly input by the user, or can be generated based on the building effect picture data and / or the pre-collected historical poster data by using a neural network model, a large language model or other models based on artificial intelligence technology.

[0068] Through the above method, the embodiment can integrate data query, knowledge base and AI technology, realize the automation of data collection, and reduce manual intervention.

[0069] Moreover, the embodiment only obtains the building effect picture data in the knowledge base owned by the system, thereby avoiding the infringement risk of using third-party materials.

[0070] Optionally, see Figure 2 In step S102, the process of determining at least one backup picture from the real estate effect picture data according to the description text can include the following steps.

[0071] S201, selecting at least one target initial picture from the multiple initial pictures contained in the real estate effect picture data according to the description text.

[0072] S202, matching a target picture style from the multiple preset picture styles according to the description text.

[0073] S203, processing at least one target initial picture according to the target picture style to obtain at least one backup picture with the target picture style.

[0074] In step S201, a feature extraction model with text understanding and image understanding capabilities can be used to extract text features of the description text and picture features of each initial picture in the real estate effect picture data. The feature extraction model can be a neural network model, a large language model, or other models.

[0075] Then the similarity of the text features and the picture features can be calculated, and the target initial picture can be selected according to the similarity, for example, selecting the initial picture with a similarity greater than a preset threshold as the target initial picture or selecting the top 10, 20 or other number of initial pictures in descending order of similarity as the target initial picture.

[0076] Optionally, when the real estate effect picture data contains a real estate three-dimensional model, a large language model can be used to analyze the description text to determine a suitable angle and position according to the description text, take a picture of the real estate three-dimensional model at the angle and position, and obtain the picture at the corresponding angle and position as the target initial picture, and then process the backup picture in the manner of S202 and S203.

[0077] In step S202, for each preset picture style pre-recorded in the database, a style label text corresponding to the picture style can be obtained, which can be input by the user when recording the picture style. In this scheme, more than 10 basic picture styles can be pre-set in the database, such as modern minimalist style, retro nostalgic style, etc.

[0078] Then the matching degree of the style label text and the style text contained in the description text can be calculated, so as to determine the style label text with the highest matching degree, and finally determine the picture style corresponding to the style label text with the highest matching degree as the target picture style.

[0079] For example, the style text is "simple style and modernized", and the matching target picture style can be modern minimalist style.

[0080] Optionally, the description text can also include property information of the real estate, such as including villas, apartments, general commercial residences, etc. In S202, the property information of the real estate can also be input into the large language model as a prompt word, so that the large language model automatically determines the target picture style matched with the property information of the real estate.

[0081] Optionally, in S202, a plurality of basic picture styles can also be determined according to the description text by the foregoing method, and the target picture style can be obtained by mixing the plurality of basic picture styles determined by using the large language model, that is, the target picture style can be obtained by fusing a plurality of basic picture styles.

[0082] In step S203, the picture style of each target initial picture can be identified. If the picture style of the target initial picture is originally the target picture style, no processing is required, and the target initial picture is directly determined as a backup picture. If the picture style of the target initial picture is not the target picture style, a pre-trained picture style conversion model is used to process the target initial picture based on the target picture style, and at least one backup picture with the target picture style is obtained.

[0083] The picture style conversion model can be any image processing model capable of converting picture styles in related technologies, and the structure and working principle thereof will not be described herein.

[0084] Optionally, the description text includes structure text, style text, and content text.

[0085] Please refer to Figure 3 According to the backup picture and the description text, the target web page file can be generated, which can include:

[0086] S301, processing the content text according to the style text to obtain a to-be-displayed text with a target style described by the style text and consistent content with the content text.

[0087] S302, according to the layout structure described by the structure text, the to-be-displayed text and the backup picture are laid out to obtain a target web page file.

[0088] The target style can include but is not limited to font, size, color, and other attributes of the to-be-displayed text. In the embodiment, the style text can be input into the large language model as a prompt word, and a text style matched with the description of the style text is selected from a plurality of preset text styles based on the large language model, and the text style is determined as the target style. Then, the content text is input into the large language model as a prompt word, and the content text is rendered based on the target style by using related text rendering technology to generate a to-be-displayed text with the same content and the target style.

[0089] As an example, the content text can be "XX building 8 discount", the generated to-be-displayed text is also "XX building 8 discount", and the to-be-displayed text has the target style regulation of A font, a size of 3 size specified by the target style, and a color of red specified by the target style.

[0090] In step S302, the structural text can be first input into the large language model as a prompt word, so that the large language model generates a layout structure of a webpage conforming to the description of the structural text, and finally a webpage editing tool in the related art is used to fill the to-be-displayed text and the backup picture into the specified position of the layout structure to obtain a target webpage file.

[0091] The target webpage file can be specifically a HyperText Markup Language (HTML) code file.

[0092] Through the above method, the embodiment can use the prompt word design of three layers of structure layer (i.e., structural text) + style layer (i.e., style text) + data layer (i.e., content text) to constrain the large language model to generate a standardized and maintainable target webpage file, so as to automatically generate the target webpage file based on artificial intelligence technology and reduce manual intervention. Moreover, by generating the target webpage file in this way, it can support dynamically inserting copyright compliant content in the webpage, such as pictures of the own knowledge base and text based on open source fonts.

[0093] Optionally, please refer to Figure 4 , based on the target webpage file, a target picture can be obtained by taking a screenshot, which can include the following steps.

[0094] S401, compliance detection is performed on the text contained in the target webpage file.

[0095] S402, in response to detecting non-compliant text that does not meet the preset text specification condition, the non-compliant text is replaced with compliant text having the same or similar semantics to obtain a replacement webpage file.

[0096] S403, a target picture is obtained by taking a screenshot in a display interface displaying the replacement webpage file.

[0097] The compliance detection refers to detecting whether the text contained in the target webpage file meets the preset text specification condition.

[0098] The content of the text specification condition is not limited. As some examples, the text specification condition can include any one or more of the following:

[0099] The font of the displayed text in the webpage is an open source font;

[0100] The displayed text in the webpage does not contain sensitive words.

[0101] The sensitive word detection module is embedded in the HTML generation stage to automatically filter the non-compliant description text and replace it with compliant content.

[0102] In S402, if the detected non-compliant text is text not using an open source font, an open source font closest to the current font of the non-compliant text can be selected from a plurality of open source fonts, and then a text using the selected open source font and having the same content is generated as a compliant text. The corresponding non-compliant text in the target web page file is replaced with the compliant text.

[0103] If the detected non-compliant text is text containing a sensitive word, the non-compliant text can be input as a prompt word into a large language model, and a compliant text having the same or similar semantics as the non-compliant text is generated by the large language model to replace the corresponding non-compliant text in the target web page file.

[0104] In some optional embodiments, it is also possible to detect whether there is a sensitive word in the description text before generating the target web page file, and if there is, replace the sensitive word with other non-sensitive words according to the above method to obtain the replaced description text, and then generate the target web page file containing no sensitive words based on the replaced description text according to the foregoing method.

[0105] Through the above method, on the one hand, the target web page file can use open source fonts to avoid the risk of infringement of using third-party font materials, and on the other hand, the sensitive words in the text can be filtered to avoid generating a target web page file containing words that violate relevant regulations.

[0106] Optionally, please refer to Figure 5 , based on the target web page file, a target picture can be obtained by taking a screenshot, which can include the following steps.

[0107] S501, obtain an adjustment instruction.

[0108] S502, adjust the web page layout of the target web page file according to the adjustment instruction to obtain an adjusted web page file, the web page layout including at least one of text size, text position, picture size and picture position.

[0109] S503, take a screenshot in a display interface displaying the adjusted web page file to obtain a target picture.

[0110] In step S501, the system can obtain the adjustment instruction in response to the user's operation, for example, a visual editing interface can be displayed based on the target web page file, and the adjustment instruction for adjusting the picture size can be obtained in response to the user's drag operation on the picture frame in the interface, and the adjustment instruction for adjusting the text position can be obtained in response to the user's drag operation on the text center in the interface.

[0111] Alternatively, the system can analyze the natural language input by the user through voice or keyboard by using a large language model to generate adjustment instructions in line with the user's intention, for example, the input natural language can be "make picture 1 a little smaller", and the generated adjustment instruction can be an adjustment instruction for reducing the size of picture 1 by a preset amplitude.

[0112] In S502, the obtained adjustment instruction can be executed by using any web page editing tool in the related art to adjust any one of the specified text size, text position, picture size and picture position in the target web page file, and the specific adjustment method is not described herein.

[0113] In actual application, the compliance detection and processing can be performed according to the method of Figure 4 , the adjustment can be performed according to the method of Figure 5 , and finally the target picture is obtained by screenshot, or the adjustment can be performed according to the method of Figure 5 , the compliance detection and processing can be performed according to the method of Figure 4 , and finally the target picture is obtained by screenshot.

[0114] Optionally, when taking a screenshot of the web page file displayed in the display interface, the Chrome browser or other browsers can be used to display the web page file, and when the Chrome browser is used to display, the Chromedp screenshot tool integrated in the Chrome browser can be used to take a screenshot. When taking a screenshot, the window resolution can be set according to the user's needs, and the display of the web page file on a specified device type can be simulated, for example, the display of the web page file on an iPhone 12 Pro can be simulated.

[0115] wherein the window resolution determines the resolution of the target picture obtained by the screenshot, for example, if the window resolution is 1920x1080, then the resolution of the target picture obtained by the screenshot is 1920x1080.

[0116] Through the above method, the browser rendering screenshot can be fully automated without human intervention.

[0117] The present application also provides a picture generation device, please refer to Figure 6 , the device can include the following units.

[0118] The obtaining unit 601 is configured to obtain picture generation information, and the picture generation information includes building effect picture data and description text.

[0119] The determining unit 602 is configured to determine at least one standby picture from the building effect picture data according to the description text.

[0120] The generating unit 603 is configured to generate a target web page file according to the standby picture and the description text.

[0121] The screenshot unit 604 is configured to perform screenshotting based on the target webpage file to obtain the target picture.

[0122] Optionally, when the screenshot unit 604 performs screenshotting based on the target webpage file to obtain the target picture, the screenshot unit 604 is specifically configured to:

[0123] perform compliance detection on the text contained in the target webpage file;

[0124] in response to detecting non-compliant text that does not satisfy the preset text specification condition, replace the non-compliant text with compliant text having the same semantics or similar semantics to obtain a replacement webpage file;

[0125] perform screenshotting in a display interface displaying the replacement webpage file to obtain the target picture.

[0126] Optionally, when the screenshot unit 604 performs screenshotting based on the target webpage file to obtain the target picture, the screenshot unit 604 is specifically configured to:

[0127] obtain an adjustment instruction;

[0128] adjust a webpage layout of the target webpage file according to the adjustment instruction to obtain an adjusted webpage file, the webpage layout including at least one of a text size, a text position, a picture size, and a picture position;

[0129] perform screenshotting in a display interface displaying the adjusted webpage file to obtain the target picture.

[0130] Optionally, the description text includes structural text, style text, and content text.

[0131] When the generation unit 603 generates the target webpage file according to the backup picture and the description text, the generation unit 603 is specifically configured to:

[0132] process the content text according to the style text to obtain to-be-displayed text having a target style described by the style text and consistent content with the content text;

[0133] layout the to-be-displayed text and the backup picture according to a layout structure described by the structural text to obtain the target webpage file.

[0134] Optionally, when the determination unit 602 determines at least one backup picture from the real estate effect picture data according to the description text, the determination unit 602 is specifically configured to:

[0135] select at least one target initial picture from a plurality of initial pictures contained in the real estate effect picture data according to the description text;

[0136] match a target picture style from a plurality of preset picture styles according to the description text;

[0137] According to the target picture style processing at least one target initial picture, at least one standby picture with the target picture style is obtained.

[0138] The working principle of the picture generation apparatus of the embodiment can be referred to the related steps of the picture generation method of the foregoing embodiments, and will not be described herein.

[0139] It should be noted that each of the embodiments in the present specification is described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The same or similar parts between the embodiments can be referred to each other.

[0140] For the convenience of description, the above system or apparatus is described in various modules or units in terms of functions. Of course, the functions of each unit can be implemented in the same or multiple software and / or hardware in the implementation of the present application.

[0141] From the above description of the embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software and the necessary general hardware platform. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes a plurality of instructions to make a computer device (which can be a personal computer, server, or network device, etc.) execute the methods described in various embodiments or some parts of the embodiments.

[0142] Finally, it should be noted that in this paper, relationship terms such as first, second, third and fourth are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitations, the element defined by the statement "including a" does not exclude the presence of another identical element in the process, method, article or device including the element.

[0143] The above is only the preferred embodiment of the present application. It should be noted that for those skilled in the art, without departing from the principle of the present application, a number of improvements and refinements can be made, which should be considered as the protection scope of the present application.

Claims

1. A picture generation method, characterized by, The method comprises the following steps: obtaining picture generation information, wherein the picture generation information comprises real estate effect picture data and description text; determining at least one backup picture from the real estate effect picture data according to the description text; generating a target web page file according to the backup picture and the description text; performing screenshot based on the target web page file to obtain a target picture.

2. The method of claim 1, wherein, The method of performing screenshot based on the target web page file to obtain a target picture comprises the following steps: performing compliance detection on the text contained in the target web page file; in response to detecting non-compliant text that does not meet the preset text specification condition, replacing the non-compliant text with compliant text having the same semantics or similar semantics to obtain a replacement web page file; performing screenshot in a display interface displaying the replacement web page file to obtain a target picture.

3. The method of claim 1, wherein, The method of performing screenshot based on the target web page file to obtain a target picture comprises the following steps: obtaining an adjustment instruction; adjusting the web page layout of the target web page file according to the adjustment instruction to obtain an adjusted web page file, wherein the web page layout comprises at least one of text size, text position, picture size and picture position; performing screenshot in a display interface displaying the adjusted web page file to obtain a target picture.

4. The method of claim 1, wherein, The description text comprises structure text, style text and content text. The method of generating a target web page file according to the backup picture and the description text comprises the following steps: processing the content text according to the style text to obtain display text having a target style described by the style text and consistent content with the content text; formatting the display text and the backup picture according to the layout structure described by the structure text to obtain a target web page file.

5. The method of claim 1, wherein, The method of determining at least one backup picture from the real estate effect picture data according to the description text comprises the following steps: selecting at least one target initial picture from a plurality of initial pictures contained in the real estate effect picture data according to the description text; matching a target picture style from a plurality of preset picture styles according to the description text; processing the at least one target initial picture according to the target picture style to obtain at least one backup picture having the target picture style.

6. An image generating apparatus characterized by comprising: The method comprises the following steps: an obtaining unit is configured to obtain picture generation information, wherein the picture generation information comprises real estate effect picture data and description text; a determining unit is configured to determine at least one backup picture from the real estate effect picture data according to the description text; a generating unit is configured to generate a target web page file according to the backup picture and the description text; a screenshot unit is configured to perform screenshot based on the target web page file to obtain a target picture.

7. The apparatus of claim 6, wherein, When the screenshot unit performs screenshot based on the target web page file to obtain a target picture, the screenshot unit is specifically configured to: perform compliance detection on the text contained in the target web page file; in response to detecting non-compliant text that does not meet the preset text specification condition, replace the non-compliant text with compliant text having the same semantics or similar semantics to obtain a replacement web page file; perform screenshot in a display interface displaying the replacement web page file to obtain a target picture.

8. The apparatus of claim 6, wherein, When the screenshot unit performs screenshot based on the target web page file to obtain a target picture, the screenshot unit is specifically configured to: obtaining an adjustment instruction; adjusting a webpage layout of the target webpage file according to the adjustment instruction, the webpage layout comprising at least one of a text size, a text position, a picture size and a picture position, to obtain an adjusted webpage file; capturing a screenshot in a display interface displaying the adjusted webpage file to obtain a target picture.

9. The apparatus of claim 6, wherein, The description text comprises a structure text, a style text and a content text. When the generating unit generates the target webpage file according to the backup picture and the description text, the generating unit is specifically configured to: process the content text according to the style text to obtain a to-be-displayed text having a target style described by the style text and consistent content with the content text; layout the to-be-displayed text and the backup picture according to a layout structure described by the structure text to obtain the target webpage file.

10. The apparatus of claim 6, wherein, When the determining unit determines at least one backup picture from the real-estate effect picture data according to the description text, the determining unit is specifically configured to: select at least one target initial picture from a plurality of initial pictures contained in the real-estate effect picture data according to the description text; match a target picture style from a plurality of preset picture styles according to the description text; process the at least one target initial picture according to the target picture style to obtain at least one backup picture having the target picture style.

Images