A method and system for improving sales efficiency based on multi-dimensional video analysis

By analyzing the live broadcast room video comment data and generating characteristic live broadcast room data as promotion videos, the problem that promotion videos in the existing technology cannot meet the shopping psychology and improve the sales effect of live broadcast goods.

CN114897585BActive Publication Date: 2024-11-29厦门蝉羽网络科技有限公司

Patent Information

Application Number
CN202210630099.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-06-06
Publication Date
2024-11-29
Estimated Expiration
2042-06-06

AI Technical Summary

Technical Problem

The live streaming promotion videos in the prior art cannot meet the shopping psychology, resulting in the problem of poor sales results.

Method used

By analyzing the comment data of the live broadcast room video, the relationship between this time and the historical data is extracted, and characteristic live broadcast room data is generated as promotional videos to satisfy the shopping psychology.

Benefits of technology

It has improved the effect of selling goods in the live broadcast room, attracted more people to enter the live broadcast room to shop, and met the psychological needs of shopping.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114897585B_ABST
    Figure CN114897585B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for improving sales efficiency based on multi-dimensional analysis of videos, comprising: acquiring live broadcast room data of this time; dividing the sales video data into different sales video segment data, and respectively obtaining corresponding comment data; for the sales video segment data, extracting relationship data of the sales video segment data and the sales video segment data of the live broadcast room data with a history of being referenced by the sales video segment data from the corresponding comment data, and generating candidate sales video segment data; calculating the importance score of each candidate sales video segment data of the sales video segment data, and synthesizing characteristic sales video segment data; according to the generation time of each sales video segment data, finally synthesizing the corresponding characteristic sales video segment data into the characteristic live broadcast room data of this time. The present invention plays the characteristic live broadcast room data as a promotional video before the next live broadcast, which can satisfy people's shopping psychology and thus attract people to enter the live broadcast room for shopping.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of live streaming with goods, and specifically relates to a method and system for improving the efficiency of live streaming with goods based on multi-dimensional analysis of videos. Background Art

[0002] The continuous development of live broadcast technology and e-commerce has given rise to a new form of commodity sales, live broadcasting with goods. Live broadcasting with goods can not only enable people to have a more comprehensive understanding of commodity information, but also enhance people's shopping experience. People can participate in discussions about commodities in the live broadcast room through real-time comments to satisfy their shopping psychology. However, the live broadcasting with goods in the existing technology has the problem of poor sales effect. One of the reasons for this problem is that the promotional video played to people before the live broadcast with goods starts cannot satisfy people's shopping psychology, thereby failing to attract people to enter the live broadcast room for shopping. Therefore, it is of great practical significance to study a method for generating promotional videos based on the analysis of live broadcast room videos to thereby improve the sales effect of live broadcast rooms. Summary of the invention

[0003] In response to the technical problems raised above, the present invention extracts the relationship between the current live broadcast room data and the historical live broadcast room data based on the comments on the live broadcast video with goods generated when people are shopping in the live broadcast room, and also generates the current feature live broadcast room data based on the relationship. The feature live broadcast room data can better meet people's shopping psychology, and therefore the feature live broadcast room data is used as a promotional video to attract people to shop in the live broadcast room.

[0004] In order to achieve the above-mentioned purpose of the invention, a method for improving sales efficiency based on multi-dimensional video analysis is provided as follows, which specifically includes the following steps:

[0005] Obtain the live broadcast room data, which includes the product-carrying video data in the live broadcast room and the comment data of the users in the live broadcast room on the product-carrying video data;

[0006] Dividing the product-carrying video data into different product-carrying video segment data, and respectively obtaining the comment data corresponding to each product-carrying video segment data;

[0007] For the goods-carrying video segment data, extract the relationship data between the goods-carrying video segment data and the goods-carrying video segment data of the pre-stored historical live broadcast room data from the corresponding comment data, and generate candidate goods-carrying video segment data for the goods-carrying video segment data according to the relationship data;

[0008] Calculate the importance score of each candidate video segment data of the video segment data, and synthesize the candidate video segment data of the ...

[0009] According to the generation time of each of the above-mentioned video segment data, the corresponding characteristic video segment data is finally synthesized into the characteristic live broadcast room data of this time, which is used to play as the promotional video of the live broadcast room before the next live broadcast of the product;

[0010] Calculating the importance score of each candidate video segment data of the video segment data includes the following steps:

[0011] Decompose the string of the comment data of the product-selling video segment data into several nouns;

[0012] A candidate video segment data of selling goods is selected, and a character string of comment data of the candidate video segment data of selling goods is decomposed into a plurality of nouns;

[0013] Compare the similarity of a noun in the comment data of the video segment data with each noun in the comment data of the candidate video segment data in step 2, calculate the importance score of each noun, and accumulate the importance score of each noun to obtain the importance score of the candidate video segment data in step 2;

[0014] Determine whether each candidate product promotion video segment has been processed in step 3. If so, proceed to the next step. Otherwise, jump to step 2 to continue execution.

[0015] Determine whether each noun in the comment data of the video segment data has been processed in step three. If so, end the step. Otherwise, select the next noun in the comment data of the video segment data, and jump to step three to continue execution.

[0016] Before comparing the similarity of two nouns, each dictionary noun is stored in advance according to the following storage structure: a noun is determined, and the superordinate noun that can be traced back from the noun is used as the superior node, the subordinate noun that can be extended from the noun is used as the subordinate node, and the synonymous noun of the noun is used as the same-level node. At the same time, the levels of different nouns increase in sequence from superior to subordinate, and the initial level is 0;

[0017] Use the following formula to compare the similarity of two nouns and calculate the importance score of the noun:

[0018] ;

[0019] in, Score the importance of the noun, , are the levels of two nouns to be compared for similarity in the same storage structure, for , The one with the smaller series.

[0020] Compared with the prior art, the beneficial effects of the present invention are at least as follows:

[0021] 1. A method for improving the efficiency of live streaming based on multi-dimensional analysis of videos of the present invention, first, obtain the live streaming data of this time; then, divide the live streaming video data into different live streaming video segment data, and obtain corresponding comment data respectively; secondly, for the live streaming video segment data, extract the relationship data of the live streaming video segment data of the live streaming video segment data with the reference history of the live streaming video segment data from the corresponding comment data, and generate candidate live streaming video segment data; thirdly, calculate the importance score of each candidate live streaming video segment data of the live streaming video segment data, and synthesize the characteristic live streaming video segment data; finally, according to the generation time of each live streaming video segment data, the corresponding characteristic live streaming video segment data is finally synthesized into the characteristic live streaming data of this time;

[0022] 2. The present invention can solve the problem in the prior art that the promotional video played to people before the start of live streaming with goods cannot satisfy people's shopping psychology and thus cannot attract people to enter the live streaming room for shopping. The present invention generates characteristic live streaming room data based on the comments about the goods-carrying video generated by people when shopping in the live streaming room, and plays it as a promotional video for the live streaming room before the next live streaming with goods starts, which can better satisfy people's shopping psychology, thereby attracting more people to enter the live streaming room for shopping, and thus improving the effect of live streaming with goods. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a flowchart of the steps of a method for improving sales efficiency based on multi-dimensional video analysis of the present invention;

[0024] Figure 2 This is a structural diagram of the composition of a system for improving sales efficiency based on multi-dimensional video analysis of the present invention. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0026] It is understood that the terms "first", "second", etc. used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, without departing from the scope of this application, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script.

[0027] References Figure 1 As shown, the present invention provides a method for improving the efficiency of bringing goods based on multi-dimensional analysis of videos, which is mainly achieved by executing the following steps:

[0028] S1. Obtain the live broadcast room data of this time, wherein the live broadcast room data includes the product-carrying video data of the live broadcast room and the comment data of the users of the live broadcast room on the product-carrying video data;

[0029] S2, dividing the above-mentioned video data into different video segment data, and respectively obtaining the above-mentioned comment data corresponding to each video segment data;

[0030] S3. For the goods-carrying video segment data, extract the relationship data between the goods-carrying video segment data and the goods-carrying video segment data of the pre-stored historical live broadcast room data from the corresponding comment data, and generate candidate goods-carrying video segment data for the goods-carrying video segment data according to the above relationship data;

[0031] S4, calculating the importance score of each candidate video segment data of the video segment data of the goods, and synthesizing the candidate video segment data of the goods with the importance score exceeding the preset importance score threshold together with the video segment data of the goods in the order of the importance score from large to small to form the feature video segment data of the goods;

[0032] S5. Based on the generation time of each of the above-mentioned video segment data for selling goods, the corresponding characteristic video segment data for selling goods is finally synthesized into the characteristic live broadcast room data of this time, which is used to be played as a promotional video for the live broadcast room before the next live broadcast for selling goods.

[0033] Furthermore, the comment data of the live broadcast room users on the sales video data in the above-mentioned step S1 includes character strings, pictures, and network links to the sales video segment data of the historical live broadcast room data.

[0034] Specifically, the inventors took into account that the comments on the sales videos generated when people are doing live shopping can fully represent people's shopping psychology. For example, people may ask in their comments about the products that have been sold in the live broadcast room in the past, and they may also ask in their comments whether the products have specific attributes, etc. Therefore, based on the comment data left by people in the live broadcast room, a promotional video is generated from the sales video data in the live broadcast room. The promotional video can well satisfy people's shopping psychology, thereby attracting more people to enter the live broadcast room for shopping, and further improving the sales effect of the live broadcast room.

[0035] Furthermore, the process of dividing the sales video data into different sales video segment data in the above-mentioned step S2 includes dividing the sales video data according to individual products in the live broadcast room, dividing the sales video data according to the product types in the live broadcast room, and dividing the sales video data according to different anchors in the live broadcast room.

[0036] Specifically, the process of dividing the sales video data into different sales video segment data is not limited to using only the above method. In actual application scenarios, it can be artificially designed as needed. The above method ensures that each sales video segment data is semantically complete, thereby ensuring that the final generated feature live broadcast room data will not have problems such as video scene incoherence when played to people, so that people have a better video viewing experience.

[0037] Furthermore, before generating candidate video segment data for selling goods in the above step S3, the following steps are also included:

[0038] Determine the historical video segment data referenced by the current video segment data from the comment data of the current video segment data, and set the depth value of the current video segment data to 0, and set the depth value of the above historical video segment data to 1;

[0039] Determine the historical video segment data with goods referenced by the historical video segment data with goods from the comment data of the historical video segment data with goods, and set the depth value of the historical video segment data with goods referenced by the historical video segment data with goods to 2;

[0040] Repeat the same method as in the previous step to continue to obtain several historical product promotion video segment data, and their depth values ​​are increased by 1 one by one.

[0041] Specifically, because the comment data of the current video segment data of selling goods may contain content about historical video segment data of selling goods, the relationship data of the current video segment data of selling goods citing the historical video segment data of selling goods can be extracted from the comment data. Similarly, the comment data of the historical video segment data of selling goods may also contain content about other historical video segment data of selling goods, and thus the relationship data of the historical video segment data of selling goods citing other historical video segment data of selling goods can be extracted from the comment data. That is, the historical video segment data of selling goods may also reference other historical video segment data of selling goods, and so on, so that several historical video segment data of selling goods with different depth values ​​can be obtained.

[0042] Furthermore, generating candidate video segment data for selling goods in the above step S3 specifically includes the following steps:

[0043] The maximum depth value interval between the current sales video segment data and the historical sales video segment data that are allowed to become candidate sales video segment data is set to d, and the maximum number of candidate sales video segment data that the current sales video segment data can have is set to k;

[0044] Determine whether the number of candidate product-carrying video segment data that the product-carrying video segment data already has is less than or equal to k. If so, proceed to the next step. Otherwise, the step ends.

[0045] Determine whether the depth value of the historical sales video segment data is less than or equal to the depth value of the current sales video segment data plus d. If so, proceed to the next step. Otherwise, jump to step five.

[0046] The historical product-selling video segment data in step 3 is used as candidate product-selling video segment data, and the process jumps to step 2 to continue.

[0047] Determine the next historical video segment data with sales, and jump to step three to continue to determine whether the next historical video segment data with sales can become a candidate video segment data with sales.

[0048] Specifically, in the above method, the maximum depth value interval between the current video segment data with goods and the historical video segment data with goods that are allowed to become candidate video segment data with goods, as well as the maximum number of candidate video segment data with goods that the current video segment data with goods can have are set, and the historical video segment data with goods that meet the above number are selected from the historical video segment data with goods whose depth value interval with the current video segment data with goods is within the above maximum depth value interval as candidate video segment data with goods. This is not only to limit the length of time of the feature video segment data with goods generated by the video segment data with goods in subsequent steps to avoid the time being too long, but also to ensure the content relevance of each candidate video segment data with goods and the video segment data with goods, to avoid the content of each candidate video segment data with goods and the video segment data with goods being quite different, which would cause content jumps in the feature video segment data generated by the video segment data with goods in subsequent steps.

[0049] Furthermore, the calculation of the importance score of each candidate video segment data of the video segment data of selling goods in the above step S4 includes the following steps:

[0050] Decompose the string of the comment data of the product-selling video segment data into several nouns;

[0051] A candidate video segment data of selling goods is selected, and a character string of comment data of the candidate video segment data of selling goods is decomposed into a plurality of nouns;

[0052] Compare the similarity of a noun in the comment data of the video segment data with each noun in the comment data of the candidate video segment data in step 2, calculate the importance score of each noun, and accumulate the importance score of each noun to obtain the importance score of the candidate video segment data in step 2;

[0053] Determine whether each candidate product promotion video segment has been processed in step 3. If so, proceed to the next step. Otherwise, jump to step 2 to continue execution.

[0054] Determine whether each noun in the comment data of the video segment data has been processed by step three. If so, end the step. Otherwise, select the next noun in the comment data of the video segment data, and jump to step three to continue execution.

[0055] Specifically, in step S3, each candidate data of the video segment data with goods has been obtained. Because the contents of each candidate data of the video segment data with goods are generally different, and the importance is also different, it is necessary to calculate the importance score of each candidate data of the video segment data with goods, so as to synthesize the candidate data of the video segment data with goods whose importance score exceeds the preset importance score threshold together with the video segment data with goods in order of importance score from large to small, and firstly decompose the character string of the comment data of the video segment data with goods and the character string of the comment data of each candidate data of the video segment data with goods into a combination of several nouns, then perform a similarity comparison between the first noun of the comment data of the video segment data with goods and each noun of the comment data of each candidate data of the video segment data with goods, and calculate the importance score of each noun, and finally perform a similarity comparison between the other nouns of the comment data of the video segment data with goods and each noun of the comment data of each candidate data of the video segment data with goods, and calculate the importance score of each noun.

[0056] Furthermore, before the similarity comparison between two nouns is performed, each dictionary noun is stored in advance according to the following storage structure: a noun is determined, and the superordinate noun that can be traced back from the above noun is used as the superior node, the subordinate noun that can be extended from the above noun is used as the subordinate node, and the synonymous noun of the above noun is used as the same-level node, and at the same time, the levels of different nouns are increased in order from superior to subordinate, and the initial level is 0;

[0057] Furthermore, the following formula is used to compare the similarity of two nouns and calculate the importance score of the nouns:

[0058] ;

[0059] in, Score the importance of the noun, , are the levels of two nouns to be compared for similarity in the same storage structure, for , The one with the smaller series.

[0060] Specifically, in this embodiment, dictionary nouns are relied upon to achieve similarity comparison between two nouns, for example, all nouns recorded in a modern Chinese dictionary, but each dictionary noun needs to be stored in advance according to a specific storage structure. First, a dictionary noun is determined, and then the superordinate noun that can be traced back from the above noun is used as the superior node, the subordinate noun that can be extended from the above noun is used as the subordinate node, and the synonymous noun of the above noun is used as the same-level node. Finally, in order from superior to subordinate, the levels of different nouns increase successively, and the initial level is 0. After storing each dictionary noun, when the two nouns that need to be compared for similarity are in the same storage structure mentioned above, the importance score of each noun in the comment data of the candidate video data with goods is calculated by the above formula. The importance score can represent the degree of similarity between each noun and the noun in the comment data of the video data with goods. That is to say, the higher the importance score, the more similar the nouns in the comment data of the candidate video data with goods are to the nouns in the comment data of the video data with goods. In particular, when the two nouns that need to be compared for similarity are synonymous nouns, the importance score is 1 and is the highest. When the two nouns that need to be compared for similarity cannot be in the same storage structure mentioned above at the same time, it means that the similarity between the two nouns is the lowest, that is, the importance score is 0. The above method can help calculate the importance score of the candidate video data with goods.

[0061] References Figure 2 As shown, the present invention also provides a system for improving sales efficiency based on multi-dimensional video analysis, which is used to implement a method for improving sales efficiency based on multi-dimensional video analysis as described above, and specifically includes the following modules:

[0062] The acquisition module is used to acquire the live broadcast room data, which includes the live broadcast room's product-carrying video data and the live broadcast room users' comment data on the product-carrying video data;

[0063] A division module, used to divide the above-mentioned video data into different video segment data, and respectively obtain the above-mentioned comment data corresponding to each video segment data;

[0064] The first analysis module is used to extract, from the corresponding comment data, the relationship data between the video segment data with goods and the video segment data with goods that references the pre-stored historical live broadcast room data with goods, and generate candidate video segment data with goods according to the relationship data;

[0065] The second analysis module is used to calculate the importance score of each candidate video segment data of the video segment data of the goods, and synthesize the candidate video segment data of the goods with an importance score exceeding a preset importance score threshold into the characteristic video segment data of the video segment data of the goods according to the order of the importance scores from large to small;

[0066] The synthesis module is used to synthesize the corresponding characteristic video segment data of selling goods into the characteristic live broadcast room data of this time according to the generation time of each of the above-mentioned video segment data of selling goods, so as to be played as the promotional video of the live broadcast room before the next live broadcast of selling goods.

[0067] In summary, the method for improving sales efficiency based on multi-dimensional analysis of videos of the present invention, first, obtains the live broadcast room data of this time; then, divides the sales video data into different sales video segment data, and obtains corresponding comment data respectively; secondly, for the sales video segment data, extracts the relationship data of the sales video segment data of the live broadcast room data with the sales video segment data citing the history from the corresponding comment data, and generates candidate sales video segment data; thirdly, calculates the importance score of each candidate sales video segment data of the sales video segment data, and synthesizes the feature sales video segment data; finally, according to the generation time of each sales video segment data, the corresponding feature sales video segment data is finally synthesized into the feature live broadcast room data of this time. The present invention can solve the problem in the prior art that the promotional video played to people before the start of live streaming with goods cannot satisfy people's shopping psychology and thus cannot attract people to enter the live streaming room for shopping. The present invention generates characteristic live streaming room data based on the comments on the goods-carrying video generated by people when shopping in the live streaming room, and plays the video as a promotional video for the live streaming room before the next live streaming with goods starts, which can better satisfy people's shopping psychology, thereby attracting more people to enter the live streaming room for shopping, and further improving the effect of live streaming with goods.

[0068] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0069] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the above-mentioned program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0070] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0071] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

[0072] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for improving sales efficiency based on multi-dimensional video analysis, characterized in that: The steps include: Obtain the live broadcast room data, which includes the product-carrying video data in the live broadcast room and the comment data of the users in the live broadcast room on the product-carrying video data; Dividing the product-carrying video data into different product-carrying video segment data, and respectively obtaining the comment data corresponding to each product-carrying video segment data; For the goods-carrying video segment data, extract the relationship data between the goods-carrying video segment data and the goods-carrying video segment data of the pre-stored historical live broadcast room data from the corresponding comment data, and generate candidate goods-carrying video segment data for the goods-carrying video segment data according to the relationship data; Calculate the importance score of each candidate video segment data of the video segment data, and synthesize the candidate video segment data of the ... According to the generation time of each of the above-mentioned video segments, the corresponding characteristic video segments are finally synthesized into the characteristic live broadcast room data, which is used to play as the promotional video of the live broadcast room before the next live broadcast; Calculating the importance score of each candidate video segment data of the video segment data includes the following steps: Decompose the string of the comment data of the product-selling video segment data into several nouns; A candidate video segment data of selling goods is selected, and a character string of comment data of the candidate video segment data of selling goods is decomposed into a plurality of nouns; Compare the similarity of a noun in the comment data of the video segment data with each noun in the comment data of the candidate video segment data in step 2, calculate the importance score of each noun, and accumulate the importance score of each noun to obtain the importance score of the candidate video segment data in step 2; Determine whether each candidate product promotion video segment has been processed in step 3. If so, proceed to the next step. Otherwise, jump to step 2 to continue execution. Determine whether each noun in the comment data of the video segment data has been processed in step three. If so, end the step. Otherwise, select the next noun in the comment data of the video segment data, and jump to step three to continue execution. Before comparing the similarity of two nouns, each dictionary noun is stored in advance according to the following storage structure: a noun is determined, and the superordinate noun that can be traced back from the noun is used as the superior node, the subordinate noun that can be extended from the noun is used as the subordinate node, and the synonymous noun of the noun is used as the same-level node. At the same time, the levels of different nouns increase in sequence from superior to subordinate, and the initial level is 0; Use the following formula to compare the similarity of two nouns and calculate the importance score of the noun: ; in, Score the importance of the noun, , are the levels of two nouns to be compared for similarity in the same storage structure, for , The one with the smaller series.

2. According to claim 1, a method for improving sales efficiency based on multi-dimensional video analysis is characterized in that: The comment data of the live broadcast room users on the product-carrying video data, including character strings, pictures, and network links of the product-carrying video segment data of the historical live broadcast room data.

3. According to claim 1, a method for improving sales efficiency based on multi-dimensional video analysis is characterized in that: The process of dividing the video data of selling goods into different video data segments of selling goods includes dividing the video data of selling goods according to individual goods in the live broadcast room, dividing the video data of selling goods according to the type of goods in the live broadcast room, and dividing the video data of selling goods according to different anchors in the live broadcast room.

4. According to claim 1, a method for improving sales efficiency based on multi-dimensional video analysis is characterized in that: Before generating candidate video segment data for selling goods, the following steps are also included: Determine the historical video segment data with goods referenced by the current video segment data with goods from the comment data of the current video segment data with goods, and set the depth value of the current video segment data with goods to 0, and set the depth value of the historical video segment data with goods to 1; Determine the historical video segment data with goods referenced by the historical video segment data with goods from the comment data of the historical video segment data with goods, and set the depth value of the historical video segment data with goods referenced by the historical video segment data with goods to 2; Repeat the same method as in the previous step to continue to obtain several historical product promotion video segment data, and their depth values ​​are increased by 1 one by one.

5. According to claim 1, a method for improving sales efficiency based on multi-dimensional video analysis is characterized in that: Generating candidate video segment data for selling goods includes the following steps: The maximum depth value interval between the current sales video segment data and the historical sales video segment data that are allowed to become candidate sales video segment data is set to d, and the maximum number of candidate sales video segment data that the current sales video segment data can have is set to k; Determine whether the number of candidate product-carrying video segment data that the product-carrying video segment data already has is less than or equal to k. If so, proceed to the next step. Otherwise, the step ends. Determine whether the depth value of the historical sales video segment data is less than or equal to the depth value of the current sales video segment data plus d. If so, proceed to the next step. Otherwise, jump to step five. The historical product-selling video segment data in step 3 is used as candidate product-selling video segment data, and the process jumps to step 2 to continue. Determine the next historical video segment data with sales, and jump to step three to continue to determine whether the next historical video segment data with sales can become a candidate video segment data with sales.

6. A system for improving sales efficiency based on multi-dimensional video analysis, used to implement the method according to any one of claims 1 to 5, characterized in that: Includes the following modules: An acquisition module is used to acquire the live broadcast room data of this time, wherein the live broadcast room data includes the goods-carrying video data of the live broadcast room, and the comment data of the live broadcast room users on the goods-carrying video data; A division module, used to divide the product-carrying video data into different product-carrying video segment data, and respectively obtain the comment data corresponding to each product-carrying video segment data; The first analysis module is used to extract, from the corresponding comment data, the relationship data between the video segment data with goods and the video segment data with goods referenced by the pre-stored historical live broadcast room data with goods, and generate candidate video segment data with goods according to the relationship data; The second analysis module is used to calculate the importance score of each candidate video segment data of the video segment data of the goods, and synthesize the candidate video segment data of the goods with an importance score exceeding a preset importance score threshold into the characteristic video segment data of the video segment data of the goods in descending order of the importance score; The synthesis module is used to synthesize the corresponding characteristic video segment data of selling goods into the characteristic live broadcast room data according to the generation time of each of the video segment data of selling goods, so as to be played as the promotional video of the live broadcast room before the next live broadcast of selling goods.

Citation Information

Patent Citations

  • Commoditymerchandising browsing value attribute evaluation method and device for shipment video

    CN112261475A

Cited By

  • Low-cost and high-accuracy e-commerce goods carrying video analysis method and system

    CN122269107A

  • A low-cost high-accuracy e-commerce goods-carrying video analysis method and system

    CN122269107B