Human-machine cooperation video generation system

Abstract

The invention discloses a human-machine cooperation video generation system, and belongs to the technical field of artificial intelligence and e-commerce video generation. The method comprises a cloud step and an end side step. The cloud step comprises a script copywriting generation step, a shooting guidance generation step and a video mixing / matching verbal skill generation step. The end side comprises the steps of collecting materials according to guidance and feeding back the materials in real time, distributing videos and uploading the videos to a video platform. The method has the advantages that millions of stores of a brand party can be assisted to quickly generate high-quality videos on a large scale, marketing is carried out on all large platforms to obtain exposure, and brand exposure and customer obtaining channels are increased.

Description

Technical Field

[0001] The present invention belongs to the technical field of artificial intelligence and e-commerce video generation, and relates to a human-machine collaborative video generation system. Background Art

[0002] The current marketing video generation solutions can be roughly divided into the following categories:

[0003] 1. Manual shooting + post-production manual editing are very costly and cannot be scaled up.

[0004] 2. Based on the tool's automatic remixing and automatic distribution, the brand headquarters or each store independently shoots videos, then uploads them to the remixing tool for remixing, generates corresponding explanation copy, and then distributes them to major platforms. Due to limited materials, remixing will cause the generated videos to be repeated and not meet the platform requirements. If store staff are required to shoot videos, because they are non-professional shooters, the videos they shoot are unqualified, the subsequent costs are very high, and the videos they shoot cannot be linked to the marketing theme, and real-time feedback cannot be obtained.

[0005] 3. Video generation based on AIGC directly generates videos by inputting text descriptions, pictures and other information. The disadvantage is that it is uncontrollable and the cost is high. Summary of the invention

[0006] The present invention aims at solving the problems in the prior art and provides a human-machine collaborative video generation system.

[0007] A human-machine collaborative video generation system includes a cloud-side step and a terminal-side step.

[0008] The cloud-based steps include generating script text, generating shooting instructions, and generating video mixing / matching scripts.

[0009] It also includes the following steps: Generate script copy based on the activity selling points, confirm it manually, and send it to the shooting director for generation.

[0010] The script copy data is passed to the video mixing / matching speech generation step,

[0011] The product library data is passed to the video mixing / matching speech generation module.

[0012] The data generated by the shooting guidance is transmitted to the module that collects materials according to the instructions and provides real-time feedback. The data that collects materials according to the instructions and provides real-time feedback is transmitted to the labeled material library module. The labeled material library information is transmitted to the video mixing / matching speech generation module. The video mixing / matching speech generation module is connected to the video distribution module and transmits data. The video distribution module is connected to the video platform and transmits data.

[0013] The end-side steps include: collecting materials according to the instructions and providing real-time feedback, distributing videos, and uploading videos to the platform.

[0014] The advantages of the present invention are: it can assist the brand's tens of millions of stores to quickly and massively generate high-quality videos, and conduct marketing on major platforms to gain exposure, thereby increasing brand exposure and customer acquisition channels.

[0015] Based on the present invention, thanks to the real-time feedback module during shooting, for unqualified materials, the problem points are fed back in real time and corrective shooting is performed, and the quality of the material itself is guaranteed. At the same time, because each end-side user can be mobilized to shoot, the amount of material can be obtained on demand, and the material is closely related to the theme.

[0016] Based on the material and product information, the sales talk and the matching video clips are generated. On the one hand, the videos generated by this method are closely related to the theme and product, on the other hand, there will be no duplication of video clips and the video quality is high. On the other hand, the sales talk generated based on the large model will not be repeated in the sales talk itself. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. As shown in the figure:

[0018] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] Embodiment 1: Figure 1 As shown, a human-machine collaborative video generation system includes cloud-side steps and terminal-side steps.

[0021] The cloud-based steps include generating script text, generating shooting instructions, and generating video mixing / matching scripts.

[0022] Generate scripts based on the selling points of the activity, confirm them manually, and send them to the shooting director for generation.

[0023] The script copy data is passed to the video mixing / matching speech generation step,

[0024] The product library data is passed to the video mixing / matching speech generation module.

[0025] The data generated by the shooting guidance is transmitted to the module that collects materials according to the instructions and provides real-time feedback. The data that collects materials according to the instructions and provides real-time feedback is transmitted to the labeled material library module. The labeled material library information is transmitted to the video mixing / matching speech generation module. The video mixing / matching speech generation module is connected to the video distribution module and transmits data. The video distribution module is connected to the video platform and transmits data.

[0026] The end-side steps include: collecting materials according to the instructions and providing real-time feedback, distributing videos, and uploading videos to the platform.

[0027] Example 2: Figure 1 As shown, a human-machine collaborative video generation system includes a centralized data management step and a distributed data collection video generation step.

[0028] The steps of video generation for distributed data collection include:

[0029] Input: The theme of the video, such as marketing theme, event theme, etc.

[0030] Output: The final generated video data with explanations is distributed to major video platforms, such as Tik Tok, Kuaishou, etc.

[0031] like Figure 1 As shown, the theme of this video generation is input into the cloud, such as a certain product promotion, a certain festival activity, etc.

[0032] Centralized data management steps include:

[0033] Step 1: Collect the large model to generate the corresponding script copy data, and confirm it through manual modification.

[0034] Based on the big model, by designing reasonable prompt words and corresponding marketing themes, such as selling point promotion, holidays, etc., for example, combined with the hotel marketing video generation of the National Day, the big model can generate a video script, such as the first step is to shoot the National Day decorations, the second step is to shoot the hotel name and introduce the brand, the third step is to shoot the hotel room and introduce the quality, the fourth step is to shoot the hotel restaurant and introduce the catering, etc.

[0035] Step 2: After the copywriting data is generated, the shooting instructions are generated based on the large model. Based on the script in the first step, the shooting instructions are generated. The second step is to shoot the hotel name in the center.

[0036] Step 3: The end-side personnel conduct on-site shooting based on the shooting instructions. During the shooting process, real-time corrections are made to check whether the lighting, viewing angle, etc. meet the requirements. If they do not meet the requirements, they will be required to reshoot.

[0037] Upload the captured material data to the cloud, tag it, and archive it. The captured pictures or video materials are randomly selected and sent to the cloud's multimodal detection model to automatically detect whether they meet the requirements of the text description in the shooting guide, such as including National Day elements and hotel names. If not, the end-side shooting personnel are prompted to reshoot, and the user is prompted through the interface to indicate where the shooting is problematic.

[0038] Step 4: Process the collected data, generate the sales pitch and the mixed video that matches the sales pitch based on the filmed material and the corresponding product attributes and selling point information. Based on the product information and video script, the big model is used to generate the corresponding sales pitch for each paragraph. Each sales pitch is matched with the video material that meets the requirements, and the timestamps are synchronized so that the sales pitch and the corresponding video material can match in time.

[0039] Step 5: The corresponding video data is distributed to video platforms such as Douyin, Kuaishou, Xiaohongshu, etc.

[0040] The present invention can be used to produce promotional videos for the e-commerce field. The brand or store manager can manage the theme of the day's marketing activities from the cloud, such as the promotion of a certain product, and then generate a video script based on the big model, and further generate shooting instructions to distribute to the mobile phone. Store staff use the software on the mobile phone to shoot based on the shooting instructions. After shooting, the material will be reviewed in real time, and unqualified ones will be re-shot. After shooting, it is uploaded to the cloud, the video mixing module performs video mixing, and generates the corresponding words for the material based on the multimodal big model such as GPT4o, and is accompanied by the corresponding background music and distributed to the end side.

[0041] The above description is only a preferred embodiment of the present invention and is 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 human-machine collaborative video generation system, characterized in that: Contains cloud-side steps and end-side steps.

2. The human-machine collaborative video generation system according to claim 1, characterized in that: The cloud-based steps include generating script text, generating shooting instructions, and generating video mixing / matching scripts.

3. The human-machine collaborative video generation system according to claim 2, characterized in that: It also includes the following steps: Generate script copy based on the activity selling points, confirm it manually, and send it to the shooting director for generation. The script copy data is passed to the video mixing / matching speech generation step, The product library data is passed to the video mixing / matching speech generation module. The data generated by the shooting guidance is transmitted to the module that collects materials according to the instructions and provides real-time feedback. The data that collects materials according to the instructions and provides real-time feedback is transmitted to the labeled material library module. The labeled material library information is transmitted to the video mixing / matching speech generation module. The video mixing / matching speech generation module is connected to the video distribution module and transmits data. The video distribution module is connected to the video platform and transmits data.

4. The human-machine collaborative video generation system according to claim 1, characterized in that: The end-side steps include: collecting materials according to the instructions and providing real-time feedback, distributing videos, and uploading videos to the platform.

Images