Advertisement video effect analysis method, device and equipment based on material position classification

Abstract

The present application relates to the technical field of image processing, solve the problem that the prior art cannot associate the advertisement with the original material, the accuracy of evaluating the effect of the material in different scenes is low, provide a kind of advertisement video effect analysis method, device and equipment based on material position classification.The method comprises: obtaining the original video material under the advertisement delivery scene;When the use frequency of the original video material is single, the total duration of the target video highlight is obtained, the material position gear is determined according to the original video material position and the total duration, and the material position gear is used to identify the relative time interval of the original video material in the target video highlight;According to the material position gear, determine the standard parameter;Calculate the standard parameter and real-time evaluation parameter, and determine the single material evaluation result.The present application can provide accurate material performance data, which helps the advertiser to optimize the use of material.

Description

[0001] This application is a divisional application of the invention patent application filed on October 24, 2024, entitled "Method, Apparatus, Equipment and Storage Medium for Evaluating the Effectiveness of Advertising Video Material Placement" with application number 202411487805.3. Technical Field

[0002] This invention relates to the field of image processing technology, and in particular to a method, apparatus, and device for analyzing the effect of advertising videos based on material location classification. Background Technology

[0003] With the rapid development of the digital advertising market, the management of advertising creatives and the analysis of campaign performance have become crucial aspects of the advertising industry. Existing advertising creative management technologies mainly include the storage, classification, and retrieval of creatives, while campaign performance analysis technologies focus on the collection and basic analysis of data after ad campaigns. These technologies typically rely on infrastructure such as cloud storage services, content management systems (CMS), and ad servers. However, existing technologies have the following limitations: creative management often relies on manual operation, which can easily lead to the reuse of creatives, version confusion and loss, and weak correlation between the final ad and the original creatives, making it impossible to accurately track the effectiveness of the creatives; the monitoring and analysis of ad performance largely depends on overall data and lacks detailed analysis of the performance of specific creatives.

[0004] Chinese patent CN104954819A discloses a method for intelligent monitoring of new media advertisements using digital watermarking encoding technology. The method includes: first, classifying and managing new media advertisements to be broadcast through an advertising material management module; then, inserting watermarks into the new media advertisements to be broadcast using a digital watermarking encoding module and assigning them unique codes; the terminal device transmitting the parsed results (containing the unique code with the digital watermark) from the user terminal to a national advertising material database via the internet or a two-way backhaul network of a broadcasting network operator; and finally, comparing the collected advertising parsing data from the user terminal with the front-end advertising material database to obtain advertising monitoring data such as which new media advertisements the user viewed and the frequency of specific advertisement exposures. Although the aforementioned patent discloses a technical solution for obtaining advertising monitoring data by comparing advertising analysis data with a material library, the correlation between the finished advertisement and the original material is not strong. Once the material is used in multiple finished advertisements or placed on different platforms, it is difficult to track its specific performance and contribution, which limits the improvement and optimization of material utilization. Furthermore, the lack of detailed analysis of the performance of specific materials makes it impossible to accurately evaluate the effect of materials in different scenarios, thus making it difficult to implement targeted optimization strategies.

[0005] Therefore, how to link the finished advertisement with the original materials and accurately evaluate the effectiveness of the materials in different scenarios is an urgent problem to be solved. Summary of the Invention

[0006] In view of this, the present invention provides a method, apparatus and equipment for analyzing the effect of advertising videos based on the location classification of materials, in order to solve the problem in the prior art that it is impossible to associate the finished advertisement with the original materials and the low accuracy of evaluating the delivery effect of materials in different scenarios.

[0007] The technical solution adopted in this invention is:

[0008] In a first aspect, the present invention provides a method for analyzing the effectiveness of advertising videos based on material location classification, the method comprising:

[0009] Obtain at least one original video clip uploaded by the target audience in the advertising scenario;

[0010] When the original video material is used only once in the ad campaign, the total duration of the target video compilation obtained by editing the original video material is obtained. Based on the position of the original video material and the total duration, the material position tier is determined. The material position tier is used to identify the relative time interval of the original video material in the target video compilation.

[0011] Based on the position and grade of the material, standard parameters are determined, wherein the standard parameters include at least the standard material play count, standard material click rate, and standard material drop rate;

[0012] The standard parameters and real-time evaluation parameters related to the evaluation of the original video footage are calculated to determine the evaluation result of a single piece of footage.

[0013] Preferably, before obtaining the total duration of the target video compilation edited from the original video footage, and determining the footage position level based on the original video footage position and total duration, the method further includes:

[0014] Scene switching detection is performed on the original video footage, and based on the scene switching detection results, the original video footage is divided into multiple scene video segments;

[0015] Using image processing technology based on frequency domain transformation, pre-set blind watermark information is embedded into each frame of each scene video segment. The video segments with embedded blind watermark information are then edited to obtain the target video compilation.

[0016] Preferably, the step of performing scene transition detection on the original video footage and dividing the original video footage into multiple scene video segments based on the scene transition detection results includes:

[0017] The original video material is decomposed into multiple frames of original video images, and the color histogram corresponding to each original video image is obtained.

[0018] The color histograms of adjacent original video images are compared to obtain the color histogram difference value.

[0019] If the color histogram difference is greater than the preset color histogram threshold, then the original video images of adjacent frames are classified into two scene video segments.

[0020] Preferably, the step of using frequency domain transform-based image processing technology to embed preset blind watermark information into each frame of each scene video segment, and then editing the scene video segments after embedding the blind watermark information to obtain the target video compilation includes:

[0021] Perform color space conversion on each frame of the video clips in each scene to obtain the target color channel;

[0022] Perform discrete cosine transform on the target color channel to convert each frame of the video clip in each scene into a scene frequency domain image;

[0023] Based on a preset key, the high-frequency region in the scene frequency domain image is modified, and the preset blind watermark information is embedded in the scene frequency domain image.

[0024] Preferably, after embedding preset blind watermark information into each frame of each scene video segment using image processing technology based on frequency domain transformation, and editing each scene video segment after embedding the blind watermark information to obtain the target video compilation, the method further includes:

[0025] Blind watermark information is extracted from the target video compilation. Based on the extracted blind watermark information, the material segment information is determined, wherein the material segment information includes the location of the original video material and the name of the original video material used in the target video compilation.

[0026] Preferably, the step of modifying the high-frequency region in the scene frequency domain image according to a preset key and embedding the preset blind watermark information into the scene frequency domain image includes:

[0027] Based on the preset blind watermark information, obtain the template size of the area to be embedded;

[0028] Based on the size of the template to be embedded, select any target region in the scene frequency domain image, wherein the size of the target region is the same as the size of the template to be embedded;

[0029] Frequency domain coefficient analysis is performed on the target region, and the sum of squares of all frequency domain coefficients in the target region is calculated as the region energy value;

[0030] The energy value of the region is compared with a preset energy threshold. If the energy value of the region is greater than the energy threshold, the target region is classified as a high-frequency region.

[0031] If the energy value of the region is less than or equal to the energy threshold, the target region is classified as a low-frequency region.

[0032] Repeatedly select any target region in the scene frequency domain image for frequency domain coefficient analysis until all regions in the scene frequency domain image are classified as high-frequency regions or low-frequency regions;

[0033] Acquire multiple frames of scene frequency domain images, and select one scene frequency domain image as the current frame image;

[0034] Using optical flow, the positions of each pixel in the high-frequency region of the current frame image are analyzed and compared with the positions of each pixel in the previous frame image to determine the optical flow field corresponding to each pixel.

[0035] The mean value of the optical flow field corresponding to each pixel is calculated to determine the mean value of the optical flow in the high-frequency region of the current frame image.

[0036] When the mean optical flow is less than the preset optical flow threshold, the high-frequency region in the current frame image is modified, and the preset blind watermark information is embedded in the scene frequency domain image.

[0037] Repeatedly select a scene frequency domain image as the current frame image to embed blind watermark information until the preset blind watermark information is embedded in all scene frequency domain images.

[0038] Preferably, before calculating the standard parameters and real-time evaluation parameters related to the evaluation of the original video material to determine the single-use material evaluation result when the original video material is used only once in the advertisement, the method further includes:

[0039] Based on the information of the material fragments, the pre-collected delivery data is statistically analyzed at the second level to determine the real-time evaluation parameters related to the evaluation of the original video material and the number of times the original video material is used in the advertising delivery. The real-time evaluation parameters include at least: real-time material play count, real-time material click-through rate, and real-time material churn rate.

[0040] Secondly, the present invention provides an advertising video effect analysis device based on material location classification, the device comprising:

[0041] The original video material acquisition module is used to acquire at least one original video material uploaded by the target object in the advertising scenario;

[0042] The material location grading module is used to obtain the total duration of the target video compilation obtained by editing the original video material when the original video material is used only once in the ad campaign, and to determine the material location grading based on the original video material location and the total duration.

[0043] The standard parameter determination module is used to determine standard parameters based on the position and grade of the material. The standard parameters include at least the standard material play count, standard material click rate, and standard material drop rate.

[0044] The material evaluation module is used to calculate the standard parameters and real-time evaluation parameters related to the evaluation of the original video material to determine the evaluation result of a single material.

[0045] Thirdly, embodiments of the present invention also provide an electronic device, including: at least one processor, at least one memory, and computer program instructions stored in the memory, which, when executed by the processor, implement the method of the first aspect described above.

[0046] Fourthly, embodiments of the present invention also provide a storage medium storing computer program instructions thereon, which, when executed by a processor, implement the method of the first aspect described above.

[0047] In summary, the beneficial effects of the present invention are as follows:

[0048] The present invention provides a method, apparatus, and device for analyzing the effectiveness of advertising videos based on material location classification. The method includes: acquiring at least one original video material uploaded by a target object in an advertising scenario; when the original video material is used only once in the advertising campaign, acquiring the total duration of a target video compilation edited from the original video material; determining the material location tier based on the original video material location and the total duration, wherein the material location tier is used to identify the relative time interval of the original video material in the target video compilation; determining standard parameters based on the material location tier, wherein the standard parameters include at least the standard material play count, standard material click-through rate, and standard material churn rate; and calculating the standard parameters and real-time evaluation parameters related to the evaluation of the original video material to determine the single material evaluation result. This invention effectively establishes a correlation between the final advertisement (i.e., the target video compilation) and the original video materials by introducing the key intermediate variable of material position tier, thereby enabling the evaluation of material performance in different deployment scenarios. Specifically, it first acquires the original video materials uploaded by the target audience and identifies whether they are used only once in the advertisement, ensuring the uniqueness and specificity of the analysis results. Then, by combining the total duration of the edited target video compilation with the start and end positions of the original materials within the compilation, its relative position range, i.e., the material position tier, is calculated. This tier standardizes the time period of material appearance in different final advertisements, avoiding evaluation bias caused by different playback timings. Based on this, standard parameters corresponding to this position tier, such as play count, click-through rate, and churn rate, are preset to construct a unified evaluation benchmark. Finally, these standard parameters are compared and analyzed with the real-time evaluation parameters generated by the materials during actual deployment to calculate the single-use material evaluation result in the specific scenario. Through this series of steps, the solution enables the tracing back from the final product to the original material, and attributes the effect based on the specific position of the material in the final product, thereby scientifically and accurately evaluating the value and performance of the material in different placement environments. Attached Figure Description

[0049] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments of the present invention will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, and these are all within the protection scope of the present invention.

[0050] Figure 1 This is a schematic diagram illustrating the overall workflow of the advertising video effect analysis method based on material location classification in Embodiment 1 of the present invention;

[0051] Figure 2 This is a schematic diagram of the process for scene switching detection of the original video material in Embodiment 1 of the present invention;

[0052] Figure 3 This is a schematic diagram of the process for determining the color histogram threshold in Embodiment 1 of the present invention;

[0053] Figure 4 This is a schematic diagram of the process of embedding preset blind watermark information into each frame of video clips in each scene in Embodiment 1 of the present invention.

[0054] Figure 5 This is a schematic diagram of the process for extracting blind watermark information from a target video compilation obtained by editing video clips from various scenes after embedding blind watermark information in Embodiment 1 of the present invention.

[0055] Figure 6 This is a schematic diagram of the process for evaluating each material fragment in Embodiment 1 of the present invention;

[0056] Figure 7 This is a flowchart illustrating the process of determining the evaluation result of a single material in Embodiment 1 of the present invention;

[0057] Figure 8 This is a structural block diagram of the advertising video effect analysis device based on material location classification in Embodiment 2 of the present invention;

[0058] Figure 9 This is a schematic diagram of the electronic device in Embodiment 3 of the present invention. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. It should be noted that, in this document, relational terms such as "first" and "second" are used merely 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 these entities or operations. In the description of the present invention, it should be understood that the terms "center," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, the element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. Where there is no conflict, embodiments of the present invention and the various features thereof can be combined with each other, all of which are within the scope of protection of the present invention.

[0060] Example 1

[0061] Please see Figure 1 Embodiment 1 of the present invention discloses a method for analyzing the effect of advertising videos based on material location classification, the method comprising:

[0062] S1: Obtain at least one original video clip uploaded by the target object in the advertising scenario;

[0063] Specifically, acquiring the original video footage uploaded by the target audience in an advertising scenario involves first receiving the video files uploaded by users and using them as initial input data. The original video footage contains various types of content, such as product advertisements, brand promotions, film and television clips, and user-generated content (UGC). It covers different scenes, themes, and styles, exhibiting diversity and complexity. Further processing of this original video footage lays the foundation for subsequent clip identification and evaluation, ensuring accurate extraction and analysis of the usage and effectiveness of each clip during editing and deployment.

[0064] S2: Perform scene switching detection on the original video material, and based on the scene switching detection results, divide the original video material into multiple scene video segments;

[0065] Specifically, scene transition detection of the original video footage involves using image processing and computer vision technologies to automatically identify transition points between different scenes in the video, such as the moment of switching from one scene to another. Based on these scene transition detection results, the original video footage is segmented into multiple scene video segments, each representing a continuous visual and content unit. This process not only accurately locates the start and end positions of each scene but also provides a clear segmentation basis for subsequent blind watermark embedding and material management. This ensures that each segment can be independently and accurately tracked and analyzed in subsequent editing, delivery, and performance evaluation. In this way, by refining the processing of video footage, each scene segment can be managed and utilized more efficiently, improving the overall accuracy of advertising delivery and the precision of performance evaluation.

[0066] In one embodiment, please refer to Figure 2 S2 includes:

[0067] S21: Decompose the original video material into multiple frames of original video images and obtain the color histogram corresponding to each original video image;

[0068] Specifically, the first step in scene transition detection is to decompose the original video footage into multiple frames. First, the continuous original video footage is broken down into individual original video images. A video per second typically contains multiple frames (e.g., 30 frames per second). Each frame represents a single instant in the video. Next, a color histogram is calculated for each frame. A color histogram is a statistical chart representing the number of pixels of each color in an image. By calculating the color histogram (denoted as H_i) for each frame, the color distribution characteristics of each frame can be obtained. These characteristics are used in subsequent scene transition detection.

[0069] S22: Calculate the difference between the color histograms of adjacent original video images to obtain the color histogram difference value;

[0070] Specifically, after acquiring the color histogram of each frame of the original video image, the color histogram difference between adjacent frames is calculated. This is achieved by calculating the difference between the color histograms of adjacent frames. Common methods include relative entropy (KL divergence) or L2 norm. For frames F_i and F_{i+1}, the color histogram difference D_i = d(H_i, H_{i+1}) is calculated, where d can be either KL divergence or L2 norm. In this way, the degree of change in color distribution between adjacent frames is quantified. The calculation of the color histogram difference is a crucial foundation for scene transition detection because scene transitions are usually accompanied by significant changes in image color.

[0071] S23: If the color histogram difference value is greater than the preset color histogram threshold, then the original video images of adjacent frames are classified into two scene video segments.

[0072] Specifically, to determine whether a scene switch exists, the calculated color histogram difference value D_i of each frame is compared with a preset color histogram threshold T_i. To improve the robustness of the detection, an adaptive threshold is used, which is dynamically adjusted according to the changes in the current frame. If the color histogram difference value D_i of a frame is greater than the adaptive threshold T_i, then a scene switch is considered to exist between frames F_i and F_{i+1}. In this case, the original video images of adjacent frames are segmented into two different scene video segments, thereby achieving scene segmentation of the video material. This method ensures that each scene video segment has consistent visual features and is significantly different from other scenes.

[0073] In one embodiment, please refer to Figure 3 The preset color histogram threshold is determined through the following steps:

[0074] S2301: Obtain the current frame image from the original video images of adjacent frames;

[0075] Specifically, during the scene switching detection process, the original video footage is processed frame by frame. First, the current frame image in the adjacent original video images is obtained, which is the frame currently being processed. This frame image is the target of the current analysis. By calculating the color histogram difference between it and the previous frame image, it can be determined whether a scene switching has occurred.

[0076] S2302: Based on the current frame image, determine the set of color histograms of all frame images preceding the current frame image;

[0077] Specifically, after acquiring the current frame image, a set of color histograms from all previous frames is collected. This set contains the color histogram (H_1, H_2, ..., H_{i-1}) for each frame from the beginning of the video to the current frame. These histograms represent the color distribution of each frame image. By analyzing this data, we can understand the color changes in the video over the previous few frames. This information is crucial for calculating the color trend of the current frame and for subsequent threshold adjustments.

[0078] S2303: Based on the set of color histograms, and in combination with preset smoothing parameters and an initial threshold, determine the color histogram threshold.

[0079] Specifically, to enhance detection robustness, an adaptive threshold is used to determine the presence of scene transitions. This is achieved by dynamically adjusting the threshold (T_i) of the current frame based on the color histogram set, combined with a preset smoothing parameter (α) and an initial threshold (T). The specific calculation method is T_i = αT + (1-α)M_i, where M_i is the average difference between the previous n frames, expressed as M_i = (1 / n)Σ_{j=i-n+1}^{i}D_j. This method dynamically adjusts the threshold based on changes in the current frame, making scene transition detection more sensitive and accurate. If the color histogram difference D_i between the current frame and the previous frame is greater than this adaptive threshold T_i, the system determines that a scene transition exists between these two frames. This adaptive threshold setting ensures flexible adjustment when processing different video materials, improving the robustness and accuracy of detection.

[0080] In one embodiment, S23 further includes:

[0081] S231: If the color histogram difference value is greater than the preset color histogram threshold, obtain the adjacent audio frame corresponding to the original video image of the adjacent frame.

[0082] Specifically, when the color histogram difference between adjacent video frames exceeds a preset color histogram threshold, it is preliminarily determined that a scene change may have occurred. At this point, in addition to relying on image information, it is also necessary to combine audio information for further verification. Therefore, audio frames corresponding to these two video frames are extracted simultaneously. An audio frame refers to a segment of audio data aligned with the video frame on the time axis. By acquiring these audio frames, the changes in the audio signal are analyzed to confirm whether a scene change has actually occurred. This process ensures the synchronous processing of video and audio information, laying the foundation for subsequent multimodal analysis.

[0083] S232: Using a preset audio feature extraction algorithm, perform audio analysis on the adjacent audio frames to determine the amplitude of audio feature changes between adjacent audio frames;

[0084] Specifically, after acquiring the audio frames, the adjacent audio frames are analyzed, and audio feature extraction algorithms are used to detect changes in the audio. The audio feature extraction algorithms include spectrum analysis, pitch change detection, audio energy analysis, frequency band energy change, and audio event detection technology. Using the preset audio feature extraction algorithms, significant changes in the audio are identified, such as background music switching, the start or end of a dialogue, or sudden changes in ambient sound. By comparing the features of adjacent audio frames, the amplitude of audio feature changes is calculated. This amplitude of change can reflect the intensity of changes in audio content in a short period of time, thereby providing an auxiliary basis for judging scene switching. If the change in audio features is large, it means that the audio environment has also changed significantly in addition to the visual scene, which further supports the judgment of scene switching.

[0085] S233: Input the original video images and audio frames of adjacent frames into a pre-trained multimodal large model, and output scene-related text content information;

[0086] Specifically, adjacent frames of raw video images are input into a pre-trained multimodal large model for processing. This multimodal large model can comprehensively process visual and audio information and automatically generate scene-related text descriptions or extract text information from the scene using deep learning technology. This step includes optical character recognition (OCR) to detect and extract text content in video frames, such as logos, subtitles, or written information within the scene. In addition, the multimodal large model can also analyze the combination of visual and audio features of the scene to generate descriptive text information. This text information may include explanations of scene content, capture of semantic changes, or even transcriptions of detected dialogue or important sounds. By generating or extracting this text information, the scene content can be further understood at the semantic level, providing richer information for scene transitions.

[0087] S234: Based on preset weights, perform weighted analysis on the change amplitude of the audio features and the text content, and determine whether to divide the original video images of adjacent frames into two scene video segments based on the weighted analysis results.

[0088] Specifically, after extracting audio and text information, a weighted analysis is performed on the changes in audio features and text content based on preset weights. The purpose of this weighted analysis is to integrate multimodal information and make more accurate scene switching judgments. The preset weights can be adjusted according to the characteristics of the video content and the analysis objectives. For example, for music videos, the weight of audio features may be higher, while for videos with rich dialogue scenes, the weight of text content may be higher. Through this weighted analysis, audio, text, and image information are combined, comprehensively considering the changes between them. If the weighted analysis results show a significant change, a scene switching is determined, and adjacent video frames are segmented into two different scene video segments. In this way, not only visual features are considered, but audio and text information are also fully utilized, ensuring more accurate and reliable scene switching detection.

[0089] S3: Using image processing technology based on frequency domain transformation, the preset blind watermark information is embedded into each frame of each scene video segment, and the video segments of each scene after embedding the blind watermark information are edited to obtain the target video compilation.

[0090] In one embodiment, please refer to Figure 4 S3 includes:

[0091] S31: Perform color space conversion on each frame of the video clip in each scene to obtain the target color channel;

[0092] Specifically, each frame of the video clips in each scene is converted from the RGB color space to the YUV color space. The YUV color space consists of three channels: Y channel (luminance), U channel (chrominance), and V channel (saturation). The converted YUV image can more effectively separate color and luminance information, allowing us to manipulate the U and V channels without affecting the overall visual quality. Through this conversion, the target color channels (U and V) of each frame are obtained, laying the foundation for subsequent blind watermark embedding.

[0093] S32: Perform discrete cosine transform on the target color channel to convert each frame image in each scene video clip into a scene frequency domain image;

[0094] Specifically, after obtaining the target color channels, a Discrete Cosine Transform (DCT) is performed on these channels. DCT is a frequency domain transformation method that converts an image from the spatial domain to the frequency domain. Through DCT, the image is represented as a set of frequency components, where low-frequency components contain the main image information, while high-frequency components contain less detailed information. This transformation allows us to process the image in the frequency domain to hide blind watermark information while minimizing the impact on the visual quality of the image. After DCT processing, the U and V channels of each frame are converted into a frequency domain image, displaying the weights of each frequency component.

[0095] S33: Based on the preset key, modify the high-frequency region in the scene frequency domain image and embed the preset blind watermark information into the scene frequency domain image.

[0096] Specifically, a preset key is used to determine the exact location for embedding blind watermark information in the frequency domain image. This key is a set of parameters used for localization, ensuring that the watermark embedding location is consistent and traceable across all frames. High-frequency regions in the scene's frequency domain image are selected for modification because these regions contain less visual information, minimizing the impact on the overall visual effect of the image. By embedding blind watermark information (i.e., a unique ID) in these high-frequency regions, each frame contains identifiable hidden information. This process involves converting the frequency domain image back to the spatial domain image using inverse IDCT, ultimately completing the embedding of the blind watermark information. In this way, each frame carries invisible identification information, ensuring that the footage can be accurately identified and tracked when used in a final cut.

[0097] In one embodiment, S33 includes:

[0098] S331: Obtain the template size of the area to be embedded based on the preset blind watermark information;

[0099] Specifically, the first step is to define the information content of the blind watermark, such as text, images, or other forms of data. Based on the complexity and information content of the watermark, a suitable region template size needs to be calculated. For example, if the watermark is a short piece of text, an 8x8 or 16x16 template size is sufficient, while if the watermark is an image or larger data, a larger area is required. The choice of template size affects the clarity and visibility of the watermark; too large a template will significantly impact the image after embedding, while too small a template will fail to convey the information completely. Therefore, a reasonable template size can improve the concealment and transmission efficiency of the watermark.

[0100] S332: Select any target region in the scene frequency domain image according to the size of the region template to be embedded, wherein the size of the target region is the same as the size of the region template to be embedded;

[0101] Specifically, a target region of the same size as the template is selected from the scene's frequency domain image. The selection method can include random selection or selection based on image features. For example, if a static scene's frequency domain image is selected, an 8x8 region can be randomly chosen. In practical applications, the selection of the target region takes into account the image's content characteristics, such as selecting a relatively flat background to ensure the watermark's concealment. If the target region happens to be located in a part with minimal texture variation, the visual effect after embedding the watermark will be better.

[0102] S333: Perform frequency domain coefficient analysis on the target region and calculate the sum of squares of all frequency domain coefficients in the target region as the region energy value;

[0103] Specifically, frequency domain coefficient analysis is performed on the selected target region. By calculating the sum of the squares of all frequency domain coefficients within the region, an energy value reflecting the intensity of that region is obtained. For example, when analyzing an 8x8 DCT coefficient matrix, the sum of the squares of each frequency domain coefficient in the matrix yields the energy value of the region. This value indicates the detail and intensity of the region and is typically used to assess its suitability for watermarking. For instance, if the energy value of a region is 50, it indicates that the region is relatively weak.

[0104] S334: Compare the energy value of the region with a preset energy threshold. If the energy value of the region is greater than the energy threshold, classify the target region as a high-frequency region.

[0105] Specifically, the calculated region energy value is compared with a preset energy threshold. If the region energy value is greater than the threshold, it is marked as a high-frequency region. For example, assuming the preset energy threshold is 40, if the energy value of a target region is 45, it indicates that the region has strong detail and is suitable for carrying a watermark. Conversely, if the energy value is lower than the threshold, it is not suitable for watermark embedding. Through this classification, the system can effectively identify which regions can effectively carry watermark information without compromising the overall image quality.

[0106] S335: If the energy value of the region is less than or equal to the energy threshold, the target region is classified as a low-frequency region;

[0107] Specifically, when the energy value of a region is less than or equal to a preset energy threshold, the region is marked as a low-frequency region. Low-frequency regions typically contain fewer detail variations and are flatter, making them suitable for embedding watermarks without causing significant visual interference. For example, if the energy value is 35, it indicates that the region is relatively stable and suitable for carrying watermarks. Through such classification, a wise choice is made between high-frequency and low-frequency regions to ensure the visual concealment of the watermark.

[0108] S336: Repeat steps S332 to S335 until all regions in the scene frequency domain image are classified as high-frequency regions or low-frequency regions.

[0109] Specifically, a comprehensive analysis of the entire scene's frequency domain image is performed. By repeating steps S332 to S335, all target regions are evaluated one by one. The purpose of this is to ensure that each region is classified so that subsequent watermark embedding strategies can be based on comprehensive data. For example, assuming an image has multiple 8x8 regions, each region is analyzed individually until all regions are labeled as high-frequency or low-frequency. This process ensures a comprehensive evaluation of the image, maximizing the selection of appropriate regions for watermark embedding.

[0110] S337: Acquire multiple frames of scene frequency domain images, and select one scene frequency domain image as the current frame image;

[0111] Specifically, the current frame image is obtained from multiple scene frequency domain images. Multiple frames provide dynamic background information for watermark embedding, which helps to improve the robustness of the watermark. For example, in a video sequence, there may be dozens of frames. Selecting one frame from these frames as the current processing object can better analyze the motion of the current frame and compare it with the previous frame, which facilitates the embedding of dynamic watermarks.

[0112] S338: Using optical flow, the positions of each pixel in the high-frequency region of the current frame image are analyzed and compared with the positions of each pixel in the previous frame image to determine the optical flow field corresponding to each pixel.

[0113] Specifically, optical flow is applied to analyze the motion of high-frequency regions in the current frame. Optical flow effectively extracts motion information by calculating changes in pixel positions between adjacent frames. For example, if a pixel in a high-frequency region of the current frame moves 5 pixels to the left, while the corresponding pixel in the previous frame remained unchanged, this motion information is recorded and an optical flow field is constructed. This process provides the foundation for subsequent watermark embedding, identifying which regions have less motion and are therefore more suitable for watermark embedding.

[0114] S339: Calculate the mean value of the optical flow field corresponding to each pixel to determine the mean value of optical flow in the high-frequency region of the current frame image;

[0115] Specifically, the mean value of all pixels in the high-frequency region of the optical flow field is calculated to determine the overall motion state of that region. For example, if the optical flow values ​​of multiple pixels in the high-frequency region are 2, 3, 4, and 5, then the mean value is 3.5. This mean value can reflect the degree of motion in that region. If the mean value is small, it means that the motion in that region is relatively small, making it suitable for watermark embedding.

[0116] S3310: When the average optical flow is less than the preset optical flow threshold, the high-frequency region in the current frame image is modified, and the preset blind watermark information is embedded in the scene frequency domain image.

[0117] Specifically, if the calculated average optical flow is lower than a preset optical flow threshold, the high-frequency region of the current frame will be modified to embed blind watermark information. For example, if the set optical flow threshold is 4 and the average is 3.5, the region will be deemed suitable for watermark embedding. In this case, the watermark information will be added to the frequency domain coefficients of the region to ensure the concealment and effectiveness of the watermark.

[0118] S3311: Repeat steps S337 to S3310 until the preset blind watermark information is embedded into all scene frequency domain images.

[0119] Specifically, finally, steps S337 to S3310 are executed repeatedly, processing each scene's frequency domain image frame by frame until all images have been embedded with the watermark. This continuous processing method ensures that the watermark information is effectively transmitted across multiple frames and improves the robustness and concealment of the watermark. By performing dynamic analysis in each frame, the consistency and stability of the watermark information throughout the entire video sequence are ensured.

[0120] S4: Extract blind watermark information from the target video compilation, and determine the material segment information based on the extracted blind watermark information. The material segment information includes at least the name and location of the original video material used in the target video compilation.

[0121] Specifically, each frame of the target video compilation is first processed. Inverse Discrete Cosine Transform (IDCT) is used to convert the image from the frequency domain back to the spatial domain, and the blind watermark information embedded in the high-frequency region is extracted. Using a preset key, the specific location and content of the blind watermark information in the image are accurately located. Based on the extracted blind watermark information, the system can identify the unique ID in each frame and determine which original footage segments were used in the target video compilation. This process ensures that even if the footage is edited, mixed, or reused in different scenes, the source and location of each footage segment can be accurately tracked, guaranteeing that the reference of footage in the final product can be perfectly identified, achieving precise data return.

[0122] In one embodiment, please refer to Figure 5 S4 includes:

[0123] S41: Decompose the target video compilation into multiple target video images, perform discrete cosine transform on each target video image, and determine the target frequency domain image;

[0124] Specifically, the target video compilation is decomposed into individual target video images, each frame representing a single moment in the video. A Discrete Cosine Transform (DCT) is applied to each frame to transform it from the spatial domain to the frequency domain, determining the target frequency domain image. Through DCT, the image is decomposed into a series of frequency components, each representing a variation in different frequencies within the image. These determined target frequency domain images reveal high-frequency and low-frequency information, providing a foundation for subsequent blind watermark extraction. This transformation enables the system to process and analyze hidden information in the image within the frequency domain.

[0125] S42: Determine the target location containing blind watermark information in the target frequency domain image based on the key;

[0126] Specifically, after acquiring the target frequency domain image, a preset key is used to determine the specific location of the blind watermark information within these images. The key is a set of predefined parameters used to identify the frequency components and locations where the blind watermark information is embedded. By referencing the key, the region containing the blind watermark information in the frequency domain image can be precisely located. This step ensures the accuracy of blind watermark extraction because only with the key can the embedded hidden information be correctly decoded. The use of the key enables the systematic location of the unique ID information hidden in each frame of the image.

[0127] S43: Based on the target location, extract and restore the blind watermark information of the target frequency domain image to determine the material fragment information.

[0128] Specifically, after determining the target location of the blind watermark information, these locations are processed to extract the hidden blind watermark information. The extraction process includes reading modification information in the high-frequency region from the frequency domain image and restoring the frequency domain information to the spatial domain through inverse discrete cosine transform (IDCT). The extracted blind watermark information contains an embedded unique ID, which corresponds one-to-one with the original material clips. Through these IDs, it is possible to identify and confirm which original material clips were used in the target video compilation, ensuring accurate tracking of the material and the precision of data feedback. In this way, detailed usage and performance data can be provided for each material clip, thereby optimizing ad placement and content management.

[0129] S5: When the target video compilation is used for advertising, each material segment is evaluated based on the pre-collected advertising data and the material segment information, and the advertising effect evaluation result is obtained by summarizing all material evaluation results.

[0130] Specifically, when the target video compilation is used for advertising, the collected campaign data is correlated and analyzed with previously extracted material clip information. The campaign data includes key metrics such as play counts, click-through rates, viewing duration, and user interaction. By matching this campaign data with the unique ID of each material clip, the performance of each material clip in different advertising scenarios can be accurately evaluated. During the evaluation process, various metrics are comprehensively considered to quantify the effectiveness of the material clips and generate a detailed evaluation report. Finally, the campaign evaluation results are output, providing advertisers with targeted optimization suggestions to help them improve the utilization efficiency and advertising effectiveness of their materials in future ad creation and campaigns.

[0131] In one embodiment, please refer to Figure 6 S5 includes:

[0132] S51: Based on the material fragment information, perform second-level statistics on the pre-collected delivery data to determine the real-time evaluation parameters related to the evaluation of the original video material and the number of times the original video material is used in the advertising delivery. The real-time evaluation parameters include at least: real-time material play count, real-time material click-through rate, and real-time material drop-off rate.

[0133] Specifically, based on the material segment information, which includes at least the name and location of the original video material used in the target video compilation, the collected advertising data is refined and statistically analyzed on a second-by-second basis. Specifically, for each original video material segment, its specific usage period in the video is first determined. For example, assuming the original video material segment is used between the 2nd and 5th seconds of the video, the data within these seconds will be the focus. For real-time material play counts, the play count at the 2nd second after the material begins use will be used as the evaluation benchmark. Next, the average churn rate per second of the original video material segment during its usage period is calculated. The real-time material churn rate refers to the proportion of viewers leaving the video in a given second. The churn rate for the 2nd to 5th seconds is calculated separately, and then the average of these four values ​​is obtained to get the material's average churn rate per second. Similarly, the click-through rate per second is also calculated. The real-time material click-through rate reflects the interaction between viewers and the advertisement during video viewing. The click-through rate for the 2nd to 5th seconds is statistically analyzed separately, and its average is calculated to obtain the material's average click-through rate per second. Through such second-level statistics, key evaluation parameters related to creative evaluation can be extracted, including play count, average churn rate, and average click-through rate. These parameters provide basic data for subsequent creative evaluation, making the evaluation process more accurate and scientific. This helps advertisers understand the specific performance of each creative in different time periods and delivery scenarios, and optimize their advertising strategies.

[0134] S52: Based on the number of uses, if the original video material is used only once, the evaluation result of the single material is determined based on the evaluation parameters and the position of the original video material.

[0135] Specifically, based on the number of uses, when the original video footage is used once, its single-use evaluation result will be calculated using previously collected evaluation parameters and the location of the original video footage. This step ensures a detailed evaluation of each footage's single use for subsequent comprehensive evaluation.

[0136] In one embodiment, please refer to Figure 7 S52 includes:

[0137] S521: Obtain the total duration of the target video compilation, and determine the media position based on the original video material location and total duration;

[0138] Specifically, based on the information of the clip, determine its specific placement within the target video compilation. This includes the clip's start and end times, as well as its exact time segment within the video. This step helps understand the clip's position within the video, providing a reference for subsequent evaluation. Based on the clip's position and the video's total length, categorize the clip's position into different tiers. For example, the video can be divided into five tiers: beginning, early, middle, late, and end. Assuming a total video length of 35 seconds, each tier would be approximately 7 seconds. The clip's center position is calculated by assuming the clip's placement is between 12 and 15.3 seconds, taking the average of 13.65, and rounding down to the nearest integer. Based on the clip's center position, calculate and determine its specific tier.

[0139] S522: Based on the material's position and grade, determine the standard parameters, wherein the standard parameters include at least the standard material's play count, standard material's click-through rate, and standard material's drop-off rate;

[0140] Specifically, after determining the specific tier of the material, standard parameters are determined based on the material's position tier. These standard parameters include at least the standard material play count, standard material click-through rate, and standard material drop-off rate. Each tier corresponds to different standard parameters. For example, five reference indicators for different positions are defined: R[0].ctr represents the standard click-through rate of the opening material, and R[4].play represents the standard play count of the closing material. The standard parameters provide a benchmark for subsequent evaluation, making the evaluation process more accurate and targeted.

[0141] S523: Calculate the standard parameters and the evaluation parameters to determine the evaluation result of the single use of the material. Specifically, the evaluation parameters of the material are compared and analyzed with the standard parameters to calculate the evaluation result of the single use of the material. The evaluation result includes scores for various indicators, comprehensively considering the performance of play count, click-through rate, and churn rate, and finally obtains a comprehensive score. The calculation formula is as follows:

[0142]

[0143] Where ε = 0.000001, x.score represents the overall score, x.play represents the number of plays of the original video clip, x.ctr represents the average click-through rate per second of the original video clip, x.lr represents the average churn rate per second of the original video clip, x.pos represents the position of the original video clip, R[x.pos].play represents the standard number of plays at the position of the original video clip, R[x.pos].ctr represents the standard click-through rate at the position of the original video clip, and R[x.pos].lr represents the standard churn rate at the position of the original video clip. This indicates the cutoff value based on the lower limit 'a' and the upper limit 'b'. This score reflects the overall effectiveness of the creative in a single use, providing advertisers with specific evaluation criteria.

[0144] S53: If the original video material is used multiple times, the weighted average of the single material evaluation results corresponding to each original video material is calculated based on the real-time material playback count to determine the multiple material evaluation results.

[0145] Specifically, when materials are used multiple times, a weighted average is calculated for the evaluation results of each use. The weighting is based on the logarithmic number of views for each use, ensuring that uses with higher view counts have a higher weight in the overall evaluation. The calculation formula is as follows:

[0146]

[0147] Here, X[i].play represents the number of plays of the original video clip when it is used for the i-th time; X[i].score represents the single-use evaluation result of the original video clip when it is used for the i-th time. In this way, a more comprehensive evaluation result can be obtained, reflecting the overall performance of the clip in multiple uses. This weighted average calculation method improves the accuracy of the evaluation and helps advertisers better understand the comprehensive effect of the clip in different placement scenarios.

[0148] Following S5, the following is also included:

[0149] S61: Obtain the ad type and platform for ad placement when targeting video highlights;

[0150] Specifically, when acquiring target video highlights for ad placement, the target platform and ad type are considered. For example, for traditional media like television, factors such as channel-changing time and viewer viewing habits are considered when evaluating content. For streaming platforms like short video platforms, user behavior data is analyzed, such as the timing and quantity of bullet comments, likes, shares, and other interactive behaviors, as well as the platform's content recommendation algorithm. Regarding ad type, Vlog ads typically emphasize natural content integration and the creator's personal style, while review ads emphasize product performance demonstrations and viewer trust. This step involves collecting and analyzing this data to provide a foundation for subsequent ad placement strategy development, ensuring that ad content and placement methods match the characteristics of the platform and ad type, thereby improving ad dissemination effectiveness and user acceptance.

[0151] S62: Based on the advertising type and the platform used for delivery, perform a second correction on the evaluation results of the materials, and use the corrected evaluation results as the results of the delivery effect analysis.

[0152] Specifically, after identifying the characteristics of the ad type and the platform, the initial material evaluation results undergo secondary correction. This process involves in-depth analysis of the ad type and platform characteristics, combining the original evaluation results with user behavior patterns and ad interaction data from the platform. For example, in Vlog ads, platform users prefer natural and personalized content, thus lowering the score for hard-sell ads and increasing the score for native content ads. In review ads, greater emphasis is placed on the clarity and credibility of product information to ensure optimal ad performance on specific platforms. Furthermore, for streaming media platforms, user interaction data such as bullet comments and likes are closely monitored, while in traditional media, the focus is on ad broadcast times and viewer viewing habits. Through this series of corrections, the final evaluation results are more accurate, better guiding subsequent ad placement and performance evaluation, ensuring that the ads achieve the expected dissemination effects.

[0153] Example 2

[0154] Please see Figure 8 Embodiment 2 of the present invention also provides an advertising video effect analysis device based on material location classification, the device comprising:

[0155] The material upload module is used to retrieve the original video materials to be uploaded;

[0156] The material segmentation module is used to perform scene switching detection on the original video material and, based on the scene switching detection results, segment the original video material into multiple scene video segments;

[0157] The blind watermark embedding module is used to embed preset blind watermark information into each frame of video clips in various scenes using image processing technology based on frequency domain transformation.

[0158] The blind watermark extraction module is used to extract blind watermark information from the target video compilation obtained by editing video clips from various scenes after embedding blind watermark information, and to determine the material clip information in the target video compilation based on the extracted blind watermark information.

[0159] The material evaluation module is used to evaluate each material segment based on pre-collected advertising data and material segment information when the target video compilation is used for advertising, and output the material evaluation results.

[0160] Specifically, the advertising video effect analysis device based on material location classification provided in this embodiment of the invention includes: a material uploading module for acquiring uploaded original video materials; a material segmentation module for performing scene switching detection on the original video materials and segmenting the original video materials into multiple scene video segments based on the scene switching detection results; a blind watermark embedding module for embedding preset blind watermark information into each frame image of each scene video segment using image processing technology based on frequency domain transformation; a blind watermark extraction module for extracting blind watermark information from the target video compilation obtained by editing each scene video segment after embedding blind watermark information and determining the material segment information in the target video compilation based on the extracted blind watermark information; and a material evaluation module for evaluating each material segment based on pre-collected advertising data and the material segment information when the target video compilation is used for advertising, and outputting material evaluation results. This device detects scene transitions in the original video footage, dividing it into multiple scene video segments. It then uses frequency domain transformation-based image processing technology to embed blind watermark information into each frame of every scene segment, ensuring each segment is uniquely identified. In the edited target video compilation, the device extracts the blind watermark information to accurately identify and track the location and usage of each original video clip. When advertising is deployed, based on pre-collected deployment data and extracted clip information, it accurately evaluates the performance of each original video clip. This not only solves the problem of weak correlation between the final advertisement and the original video footage but also provides accurate performance data through refined evaluation, helping advertisers optimize clip usage and improve advertising efficiency and effectiveness. This systematic and automated processing and analysis reduces human intervention and subjective bias, improving the accuracy and real-time nature of the evaluation.

[0161] Example 3

[0162] In addition, combined Figure 1The advertising video effect analysis method based on material location classification described in Embodiment 1 of the present invention can be implemented by an electronic device. Figure 9 A schematic diagram of the hardware structure of the electronic device provided in Embodiment 3 of the present invention is shown.

[0163] Electronic devices may include processors and memory storing computer program instructions.

[0164] Specifically, the processor may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement embodiments of the present invention.

[0165] The memory may include a large-capacity storage device for data or instructions. For example, and not limitingly, the memory may include a hard disk drive (HDD), a floppy disk drive, flash memory, an optical disk drive, a magneto-optical disk drive, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory may include removable or non-removable (or fixed) media. Where appropriate, the memory may be internal or external to a data processing device. In a particular embodiment, the memory is a non-volatile solid-state memory. In a particular embodiment, the memory includes a read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or flash memory, or a combination of two or more of these.

[0166] The processor reads and executes computer program instructions stored in the memory to implement any of the advertising video effect analysis methods based on material location classification in the above embodiments.

[0167] In one example, the electronic device may also include a communication interface and a bus. For example, Figure 9 As shown, the processor, memory, and communication interface are connected via a bus and communicate with each other.

[0168] The communication interface is mainly used to enable communication between various modules, devices, units and / or equipment in the embodiments of the present invention.

[0169] A bus, including hardware, software, or both, couples components of the device together. For example, and not limitingly, a bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, a bus may include one or more buses. While specific buses are described and illustrated in embodiments of the invention, the invention contemplates any suitable bus or interconnect.

[0170] The electronic device provided in this embodiment detects scene transitions in the original video footage, divides it into multiple scene video segments, and embeds blind watermark information into each frame of each scene segment using frequency domain transformation-based image processing technology. This ensures that each footage segment can be uniquely identified. In the edited target video compilation, the blind watermark information is extracted to accurately identify and track the location and usage of each original video footage. When advertising is deployed, the deployment effect of each original video footage is accurately evaluated based on pre-collected deployment data and extracted footage segment information. This not only solves the problem of weak correlation between the finished advertisement and the original video footage, but also provides accurate footage performance data through refined evaluation. This helps advertisers optimize footage usage and improve the efficiency and effectiveness of advertising. This systematic and automated processing and analysis reduces human intervention and subjective bias, and improves the accuracy and real-time nature of the evaluation.

[0171] Example 4

[0172] Furthermore, in conjunction with the advertising video effect analysis method based on material location classification in Embodiment 1 above, Embodiment 4 of the present invention can also provide a computer-readable storage medium for implementation. This computer-readable storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the advertising video effect analysis methods based on material location classification in the above embodiments.

[0173] The storage medium provided in this embodiment detects scene transitions in the original video footage, dividing it into multiple scene video segments. It then uses frequency domain transformation-based image processing technology to embed blind watermark information into each frame of each scene segment, ensuring that each segment is uniquely identified. In the edited target video compilation, the blind watermark information is extracted to accurately identify and track the location and usage of each original video clip. When advertising is deployed, the effectiveness of each original video clip is accurately evaluated based on pre-collected deployment data and extracted clip information. This not only solves the problem of weak correlation between the final advertisement and the original video footage but also provides accurate clip performance data through refined evaluation, helping advertisers optimize clip usage and improve the efficiency and effectiveness of advertising. This systematic and automated processing and analysis reduces human intervention and subjective bias, improving the accuracy and real-time nature of the evaluation.

[0174] In summary, the embodiments of the present invention provide a method, apparatus, and device for analyzing the effect of advertising videos based on material location classification.

[0175] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.

[0176] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this invention are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant location, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0177] It should also be noted that the exemplary embodiments mentioned in this invention describe methods or systems based on a series of steps or apparatus. However, this invention is not limited to the order of the steps described above; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0178] The above description is merely a specific embodiment of the present invention. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the protection scope of the present invention.

Claims

1. A method for analyzing the effectiveness of advertising videos based on material location classification, characterized in that, The method includes: Obtain at least one original video clip uploaded by the target audience in the advertising scenario; When the original video material is used only once in the ad campaign, the total duration of the target video compilation obtained by editing the original video material is obtained. Based on the position of the original video material and the total duration, the material position tier is determined. The material position tier is used to identify the relative time interval of the original video material in the target video compilation. Based on the position and grade of the material, standard parameters are determined, wherein the standard parameters include at least the standard material play count, standard material click rate, and standard material drop rate; The standard parameters and real-time evaluation parameters related to the evaluation of the original video footage are calculated to determine the evaluation result of a single video clip. Before obtaining the total duration of the target video compilation edited from the original video footage, and determining the footage position and tier based on the original video footage position and total duration, the process also includes: Scene switching detection is performed on the original video footage, and based on the scene switching detection results, the original video footage is divided into multiple scene video segments; Using image processing techniques based on frequency domain transformation, pre-set blind watermark information is embedded into each frame of each scene video segment. The video segments with embedded blind watermark information are then edited to obtain a target video compilation. The process of embedding preset blind watermark information into each frame of each scene video segment using frequency domain transformation-based image processing technology, and then editing the scene video segments after embedding the blind watermark information to obtain the target video compilation, further includes: Blind watermark information is extracted from the target video compilation. Based on the extracted blind watermark information, the material segment information is determined, wherein the material segment information includes the location of the original video material and the name of the original video material used in the target video compilation; Before determining the single-use evaluation result of the original video material by calculating the standard parameters and real-time evaluation parameters related to the evaluation of the original video material when the original video material is used only once in the advertisement, the process further includes: Based on the information of the material fragments, the pre-collected delivery data is statistically analyzed at the second level to determine the real-time evaluation parameters related to the evaluation of the original video material and the number of times the original video material is used in the advertising delivery. The real-time evaluation parameters include at least: real-time material play count, real-time material click-through rate, and real-time material churn rate.

2. The advertising video effect analysis method based on material location classification according to claim 1, characterized in that, The step of performing scene transition detection on the original video footage and, based on the scene transition detection results, dividing the original video footage into multiple scene video segments includes: The original video material is decomposed into multiple frames of original video images, and the color histogram corresponding to each original video image is obtained. The color histograms of adjacent original video images are compared to obtain the color histogram difference value. If the color histogram difference is greater than the preset color histogram threshold, then the original video images of adjacent frames are classified into two scene video segments.

3. The advertising video effect analysis method based on material location classification according to claim 1, characterized in that, The method utilizes frequency domain transform-based image processing technology to embed pre-set blind watermark information into each frame of each scene video segment. The resulting target video compilation is then obtained by editing the video segments after embedding the blind watermark information. Perform color space conversion on each frame of the video clips in each scene to obtain the target color channel; Perform discrete cosine transform on the target color channel to convert each frame of the video clip in each scene into a scene frequency domain image; Based on a preset key, the high-frequency region in the scene frequency domain image is modified, and the preset blind watermark information is embedded in the scene frequency domain image.

4. The advertising video effect analysis method based on material location classification according to claim 3, characterized in that, The step of modifying the high-frequency region in the scene frequency domain image according to a preset key and embedding preset blind watermark information into the scene frequency domain image includes: Based on the preset blind watermark information, obtain the template size of the area to be embedded; Based on the size of the template to be embedded, select any target region in the scene frequency domain image, wherein the size of the target region is the same as the size of the template to be embedded; Frequency domain coefficient analysis is performed on the target region, and the sum of squares of all frequency domain coefficients in the target region is calculated as the region energy value; The energy value of the region is compared with a preset energy threshold. If the energy value of the region is greater than the energy threshold, the target region is classified as a high-frequency region. If the energy value of the region is less than or equal to the energy threshold, the target region is classified as a low-frequency region. Repeatedly select any target region in the scene frequency domain image for frequency domain coefficient analysis until all regions in the scene frequency domain image are classified as high-frequency regions or low-frequency regions; Acquire multiple frames of scene frequency domain images, and select one scene frequency domain image as the current frame image; Using optical flow, the positions of each pixel in the high-frequency region of the current frame image are analyzed and compared with the positions of each pixel in the previous frame image to determine the optical flow field corresponding to each pixel. The mean value of the optical flow field corresponding to each pixel is calculated to determine the mean value of the optical flow in the high-frequency region of the current frame image. When the mean optical flow is less than the preset optical flow threshold, the high-frequency region in the current frame image is modified, and the preset blind watermark information is embedded in the scene frequency domain image. Repeatedly select a scene frequency domain image as the current frame image to embed blind watermark information until the preset blind watermark information is embedded in all scene frequency domain images.

5. An advertising video effect analysis device based on material location classification, characterized in that, The device includes: The original video material acquisition module is used to acquire at least one original video material uploaded by the target object in the advertising scenario; The material location grading module is used to obtain the total duration of the target video compilation obtained by editing the original video material when the original video material is used only once in the ad campaign, and to determine the material location grading based on the original video material location and the total duration. The standard parameter determination module is used to determine standard parameters based on the position and grade of the material. The standard parameters include at least the standard material play count, standard material click rate, and standard material drop rate. The material evaluation module is used to calculate the standard parameters and real-time evaluation parameters related to the evaluation of the original video material, and to determine the evaluation result of a single material. Before obtaining the total duration of the target video compilation edited from the original video footage, and determining the footage position and tier based on the original video footage position and total duration, the process also includes: Scene switching detection is performed on the original video footage, and based on the scene switching detection results, the original video footage is divided into multiple scene video segments; Using image processing techniques based on frequency domain transformation, pre-set blind watermark information is embedded into each frame of each scene video segment. The video segments with embedded blind watermark information are then edited to obtain a target video compilation. The process of embedding preset blind watermark information into each frame of each scene video segment using frequency domain transformation-based image processing technology, and then editing the scene video segments after embedding the blind watermark information to obtain the target video compilation, further includes: Blind watermark information is extracted from the target video compilation. Based on the extracted blind watermark information, the material segment information is determined, wherein the material segment information includes the location of the original video material and the name of the original video material used in the target video compilation; Before determining the single-use evaluation result of the original video material by calculating the standard parameters and real-time evaluation parameters related to the evaluation of the original video material when the original video material is used only once in the advertisement, the process further includes: Based on the information of the material fragments, the pre-collected delivery data is statistically analyzed at the second level to determine the real-time evaluation parameters related to the evaluation of the original video material and the number of times the original video material is used in the advertising delivery. The real-time evaluation parameters include at least: real-time material play count, real-time material click-through rate, and real-time material churn rate.

6. An electronic device, characterized in that, include: At least one processor, at least one memory, and computer program instructions stored in the memory, which, when executed by the processor, implement the method as described in any one of claims 1-4.

7. A storage medium storing computer program instructions thereon, characterized in that, The method as described in any one of claims 1-4 is implemented when the computer program instructions are executed by the processor.

Images