A digital evaluation method for film and television videos

Abstract

The present invention relates to the field of video evaluation, and in particular to a digital evaluation method for film and television videos, comprising: extracting audio and video evaluation segments for a target evaluation video, and detecting a communication complexity coefficient and a reference communication overlap coefficient of each audio and video evaluation segment; determining the category of the corresponding audio and video evaluation segment according to the communication complexity coefficient and the reference communication overlap coefficient, and determining a segment evaluation strategy for the corresponding audio and video evaluation segment according to the category of each audio and video evaluation segment; under the condition that the evaluation is completed, determining a segment focus coefficient according to an interactive effective parameter and a synchronization evaluation coefficient of each audio and video evaluation segment; under the condition that the focus analysis is completed, determining an execution compression strategy for the target evaluation video according to the proportion of the focus segments, and the execution compression strategy is to compress the target evaluation video by adopting a dynamic execution strategy or a static execution strategy. The present invention improves the accuracy of the judgment result on the audio and video synchronization of the video, thereby reducing the influence of the compression process.

Description

Technical Field

[0001] The present invention relates to the field of video evaluation, and in particular to a digital evaluation method for film and television videos. Background Art

[0002] With the development of science and technology, the difficulty of video publishing in the audio and video industry has gradually decreased. Therefore, film and television videos produced by some producers with less video production experience are prone to quality problems, such as a lack of synchronization between audio and video. Therefore, technical personnel in this field often use audio and video evaluation systems to perform quality evaluation on film and television videos. However, for continuous pictures with multiple people and rich sound sources, audio and picture quality evaluation is difficult and it is difficult to effectively achieve accurate quality evaluation. Therefore, how to perform effective quality evaluation on continuous pictures with multiple people and rich sound sources is a technical problem that needs to be urgently solved by technical personnel in this field.

[0003] Chinese patent publication number CN117012228A discloses a method and device for training an evaluation model and evaluating video quality. In the method for training the evaluation model, a sample set is obtained, including a first video sample with audio and video synchronization. The first video sample is separated from the audio and video to obtain a sample image frame sequence and a corresponding sample audio frame sequence, and the sample image frame sequence and the corresponding sample audio frame sequence are input into the evaluation model for model processing. The evaluation model includes a first encoder and a second encoder. The above model processing includes encoding each sample image frame using the first encoder to obtain each sample image representation, and encoding each sample audio frame using the second encoder to obtain each sample audio representation. With the goal of minimizing the first loss, the parameters of the first encoder and the second encoder are adjusted. The first loss is positively correlated to the distance between different modes of each sample image representation and the corresponding sample audio representation, and the distance between the same modes of adjacent sample image representations in each sample image representation. However, the above scheme has the following problems: the influence of the number of sound sources in the video sample and the audio overlap on the audio and video synchronization analysis is not taken into account, so that the accuracy of the determination result of the audio and video synchronization is low. Summary of the invention

[0004] To this end, the present invention provides a digital evaluation method for film and television videos, which is used to overcome the problem that the prior art fails to determine the targeted audio and video synchronization situation based on the communication between people in the video sample and the audio repetition, resulting in low accuracy of the determination result of the audio and video synchronization situation, and further leads to the subsequent low compression quality of the video.

[0005] To achieve the above object, the present invention provides a digital evaluation method for film and television videos, comprising:

[0006] Extract audio and video evaluation segments from the target evaluation video, and detect the communication complexity coefficient and reference communication overlap coefficient of each audio and video evaluation segment;

[0007] Determine the category of the audio-visual evaluation paragraph according to the communication complexity coefficient and the reference communication coincidence coefficient, and determine the corresponding paragraph evaluation strategy according to the category of each audio-visual evaluation paragraph;

[0008] If an audio-visual evaluation paragraph is a type of evaluation paragraph, a feature matching method is used to determine a synchronization evaluation coefficient, and the synchronization evaluation coefficient of the corresponding audio-visual evaluation paragraph is determined according to the number of feature matching paragraphs;

[0009] If the audio and video evaluation segment is a second-class evaluation segment, the synchronization evaluation coefficient is determined by using a time sequence matching method, and the synchronization evaluation coefficient is determined according to a time difference reference value between each audio segment and its corresponding video segment;

[0010] Under the evaluation completion condition, the segment focus coefficient is determined according to the interactive effective parameters and synchronous evaluation coefficients of each audio and video evaluation segment. The evaluation completion condition is that the synchronous evaluation coefficients of each audio and video evaluation segment in the target evaluation video are all determined and completed;

[0011] Under the condition of key analysis completion, the execution compression strategy of the target evaluation video is determined according to the proportion of key paragraphs. The execution compression strategy is to use a dynamic execution strategy or a static execution strategy to compress the target evaluation video. The condition of key analysis completion is that all key paragraphs in the target evaluation video are determined to be completed.

[0012] Further, the category of the audio-visual evaluation paragraph is determined according to the communication complexity coefficient and the reference communication coincidence coefficient;

[0013] If the communication complexity coefficient of an audio-visual evaluation paragraph is greater than the preset communication complexity coefficient or the reference communication coincidence coefficient is greater than the preset communication coincidence coefficient, the audio-visual evaluation paragraph is determined to be a first-class evaluation paragraph;

[0014] If the communication complexity coefficient of an audio-visual evaluation segment is less than or equal to the communication complexity coefficient and the reference communication coincidence coefficient is less than or equal to the preset communication coincidence coefficient, the audio-visual evaluation segment is determined to be a second-class evaluation segment;

[0015] The communication complexity coefficient is determined according to the number of communicators in the audio-visual evaluation segment and the reference communication frequency.

[0016] Furthermore, the reference communication overlap coefficient corresponding to any audio-visual evaluation section is the average value of the communication overlap coefficients of each speech section in the audio-visual evaluation section;

[0017] Separating the audio at the time corresponding to the audio-visual evaluation segment to obtain a number of audio segments;

[0018] For any speech segment, the communication overlap coefficient of the speech segment is determined according to the number of overlapping speech segments and the proportion of overlapping segments.

[0019] Further, determining a paragraph evaluation strategy according to the category of each audio-visual evaluation paragraph;

[0020] For any audio and video evaluation paragraph,

[0021] If the audio-visual evaluation segment is a first-class evaluation segment, a synchronous evaluation coefficient of the audio-visual evaluation segment is determined by a feature matching method;

[0022] If the audio-visual evaluation segment is a second-class evaluation segment, the synchronization evaluation coefficient of the audio-visual evaluation segment is determined by using a timing matching method, and the synchronization evaluation coefficient is determined according to a time difference reference value between each voice segment and its corresponding video segment.

[0023] Furthermore, when determining the synchronization evaluation coefficient of any audio-visual evaluation segment by using a feature matching method, the phase overlap coefficient of each video segment within the speech duration period corresponding to each speech segment of the audio-visual evaluation segment is detected, and the video segment whose phase overlap coefficient is greater than the preset phase overlap coefficient is recorded as the video segment to be evaluated of the corresponding speech segment;

[0024] Detect the feature matching coefficients of each video segment to be evaluated and the corresponding speech segment, and determine the reference feature matching coefficient of the corresponding speech segment according to the feature matching coefficients of each video segment to be evaluated;

[0025] The synchronous evaluation coefficient of any audio-visual evaluation segment is positively correlated with the number of feature matching segments, wherein the feature matching segment is a speech segment whose reference feature matching coefficient is greater than a preset reference feature matching coefficient, and the reference feature matching coefficient is the average value of the feature matching coefficients of each video segment to be evaluated.

[0026] Further, the feature matching coefficient of any video segment to be evaluated and the corresponding speech segment is determined according to the output rate matching degree of the video segment to be evaluated and the corresponding speech segment and the reference matching coefficient;

[0027] The characteristic matching coefficient is positively correlated with the output rate matching degree and the reference matching coefficient respectively.

[0028] Furthermore, the output rate matching degree is determined according to the standard matching difference value of the video segment to be evaluated, and the standard matching difference value is in a negative correlation relationship;

[0029] The reference matching coefficient is negatively correlated with an average value of the relevant feature interval durations of each video frame of the video segment to be evaluated.

[0030] Furthermore, under the evaluation completion condition, the interactive effective parameters corresponding to each audio and video evaluation paragraph of the target evaluation video are detected, and the paragraph focus coefficient is determined according to the interactive effective parameters and the synchronization evaluation coefficient of each audio and video evaluation paragraph, and the audio and video evaluation paragraph whose paragraph focus coefficient is greater than the preset paragraph focus coefficient is recorded as a focus paragraph;

[0031] The interaction validity parameter is determined according to the reference interaction amount and interaction richness of the corresponding synchronous evaluation paragraph.

[0032] Furthermore, under the condition of key analysis completion, the execution compression strategy of the target evaluation video is determined according to the proportion of key paragraphs.

[0033] If the proportion of key paragraphs is greater than the preset proportion of key paragraphs, a dynamic execution strategy is used to compress the target evaluation video;

[0034] If the proportion of key paragraphs is less than or equal to the preset proportion of key paragraphs, a static execution strategy is used to compress the target evaluation video.

[0035] Furthermore, if a dynamic execution strategy is used to compress the target evaluation video, the video experience parameters of some target evaluation videos that have completed compression in each video detection cycle are periodically detected, and it is determined whether to adjust the compression parameters according to the video experience parameters;

[0036] If the video experience parameter is less than or equal to the preset video experience parameter, determining to adjust the compression parameter of the next video detection cycle of the video detection cycle;

[0037] If the video experience parameter is greater than the preset video experience parameter, it is determined not to adjust the compression parameter of the next video detection cycle of the video detection cycle.

[0038] Compared with the prior art, the beneficial effect of the present invention lies in that, in the technical scheme of the present invention, the category of each audio and video evaluation paragraph is determined according to the communication complexity coefficient and the reference communication overlap coefficient of each audio and video evaluation paragraph, and the paragraph evaluation strategy is determined according to the category of the audio and video evaluation paragraph, that is, a targeted audio and video synchronization evaluation method is determined according to the complexity of the communication personnel in the audio and video evaluation paragraph and the overlapping of the voice paragraphs, so that the evaluation method of the audio and video synchronization of the video is more in line with the actual evaluation situation, ensuring the accuracy of the judgment result of the audio and video synchronization of the video, and further ensuring the influence of the video effect in the subsequent compression process.

[0039] Furthermore, in the present invention, for audio and video evaluation segments with a large AC complexity coefficient or a large reference AC overlap coefficient, a feature matching method is used to determine the synchronization evaluation coefficient of such audio and video evaluation segments. The feature matching coefficient is determined based on the matching of the content output rate of each speech segment contained therein and the video segment to be evaluated, as well as the matching of the content, and then the synchronization evaluation coefficient is determined. This avoids the judgment result of the audio and video synchronization of such audio and video evaluation segments being affected by content overlap. The present invention improves the accuracy of the judgment result of the video audio and video synchronization.

[0040] Furthermore, the present invention determines key paragraphs based on the acquired synchronization evaluation coefficients and interaction effectiveness parameters, that is, audio and video evaluation paragraphs with poor audio and video synchronization effects or more effective interactions are recorded as key paragraphs, and a targeted video compression method is determined based on the proportion of key paragraphs, so that the compression process is more in line with actual needs and avoids further impact of the compression process on the video effect.

[0041] Furthermore, in the present invention, when the key paragraphs account for a large proportion, compression quality detection is periodically performed on the compressed part of the video, and the parameters of the current compression process are adjusted according to the compression quality detection results. Since the key paragraphs account for a large proportion, the compression process may affect the audio and video synchronization effect. Therefore, by real-time detection of the compression quality of the video and timely adjustment of the compression process, the impact of the compression process on the video effect can be effectively reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A schematic diagram of a digital evaluation method for film and television videos of the present invention;

[0043] Figure 2 A flow chart of the present invention for determining a paragraph evaluation strategy according to the category of each audio and video evaluation paragraph;

[0044] Figure 3 A flowchart of the present invention for determining an execution compression strategy for a target evaluation video according to a proportion of key paragraphs;

[0045] Figure 4 The present invention is a flow chart for determining whether to adjust compression parameters according to video experience parameters. DETAILED DESCRIPTION

[0046] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0047] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0048] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0049] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0050] See also Figures 1 to 4 As shown, the present invention provides a digital evaluation method for film and television videos, comprising:

[0051] Extract audio and video evaluation segments from the target evaluation video, and detect the communication complexity coefficient and reference communication overlap coefficient of each audio and video evaluation segment;

[0052] Determine the category of the audio-visual evaluation paragraph according to the communication complexity coefficient and the reference communication coincidence coefficient, and determine the corresponding paragraph evaluation strategy according to the category of each audio-visual evaluation paragraph;

[0053] If an audio-visual evaluation paragraph is a type of evaluation paragraph, a feature matching method is used to determine a synchronization evaluation coefficient, and the synchronization evaluation coefficient of the corresponding audio-visual evaluation paragraph is determined according to the number of feature matching paragraphs;

[0054] If the audio and video evaluation segment is a second-class evaluation segment, the synchronization evaluation coefficient is determined by using a time sequence matching method, and the synchronization evaluation coefficient is determined according to a time difference reference value between each audio segment and its corresponding video segment;

[0055] Under the evaluation completion condition, the segment focus coefficient is determined according to the interactive effective parameters and synchronous evaluation coefficients of each audio and video evaluation segment. The evaluation completion condition is that the synchronous evaluation coefficients of each audio and video evaluation segment in the target evaluation video are all determined and completed;

[0056] Under the condition of key analysis completion, the execution compression strategy of the target evaluation video is determined according to the proportion of key paragraphs. The execution compression strategy is to use a dynamic execution strategy or a static execution strategy to compress the target evaluation video. The condition of key analysis completion is that all key paragraphs in the target evaluation video are determined to be completed.

[0057] The target evaluation video is a video that needs to be evaluated for audio and video synchronization to determine a specific compression method. The target evaluation video is divided into several audio and video evaluation segments of the same length. The duration of each audio and video evaluation segment can be set by the user according to actual needs. The higher the user's requirements for the lossless effect of video compression, the shorter the duration of each audio and video evaluation segment. A value of the duration of an audio and video evaluation segment is provided, and the value of the duration of the audio and video evaluation segment is 10 minutes.

[0058] Several historical evaluation records are applied in the present invention, and any historical evaluation record records the communication complexity coefficient, stage overlap coefficient, output rate matching parameter, feature similarity, reference feature matching coefficient, key paragraph ratio and video experience parameter in at least one monitoring record for the target management area, and each historical monitoring record corresponds to a qualified mark, which records whether the requirement for the lossless effect of video compression meets user needs.

[0059] Specifically, the category of the audio-visual evaluation paragraph is determined based on the communication complexity coefficient and the reference communication coincidence coefficient;

[0060] If the communication complexity coefficient of an audio-visual evaluation paragraph is greater than the preset communication complexity coefficient or the reference communication coincidence coefficient is greater than the preset communication coincidence coefficient, the audio-visual evaluation paragraph is determined to be a first-class evaluation paragraph;

[0061] If the communication complexity coefficient of an audio-visual evaluation segment is less than or equal to the communication complexity coefficient and the reference communication coincidence coefficient is less than or equal to the preset communication coincidence coefficient, the audio-visual evaluation segment is determined to be a second-class evaluation segment;

[0062] The communication complexity coefficient is determined according to the number of communicators in the audio-visual evaluation segment and the reference communication frequency.

[0063] Among them, for any audio and video evaluation section, the communication complexity coefficient is the natural logarithm of the product of the number of communicators in the audio and video evaluation section and the reference communication frequency of the audio and video evaluation section. The communicators are the persons with communication characteristics in the corresponding audio and video evaluation section. The reference communication frequency is the average communication frequency of each communicator. For a single communicator, the communication frequency = the number of changes in the communication characteristics of the communicator in the audio and video evaluation section / the duration of the communication of the communicator in the audio and video evaluation section. The duration of communication is the duration of the communication characteristics of the communicator in the audio and video evaluation section. For any communicator, if the communication characteristics collected by two adjacent video frames are inconsistent, it is recorded as a change in the communication characteristics;

[0064] For any speech segment, the speech segment that overlaps with the speech duration period of the speech segment is recorded as the overlapping speech segment, and the communication overlap coefficient of the speech segment is the average value of the cycle overlap of each overlapping speech segment. For any speech segment and any overlapping speech segment, the cycle overlap = the duration of the overlapping part of the two speech duration periods / the duration of the speech duration period of the speech segment. For any audio-visual evaluation segment, the reference audio-visual overlap coefficient is the average value of the audio-visual evaluation coefficients of each speech segment in the audio-visual evaluation segment.

[0065] The values ​​of the preset AC complexity coefficient and the preset AC overlap coefficient can be determined by the user according to the actual working scenario. For example, the user can set them according to the historical evaluation records. The higher the user's requirements for the lossless effect of video compression, the smaller the value of the preset AC complexity coefficient, and the smaller the value of the preset AC overlap coefficient. A method for determining the value of the preset AC complexity coefficient is provided, and the minimum value of the AC complexity coefficient of each type of evaluation paragraph in the historical evaluation records that meet the user's requirements for the lossless effect of video compression is recorded as the preset AC complexity coefficient. A method for determining the value of the preset AC overlap coefficient is provided, and the minimum value of the AC overlap coefficient of each type of evaluation paragraph in the historical evaluation records that meet the user's requirements for the lossless effect of video compression is recorded as the preset AC overlap coefficient.

[0066] Specifically, the reference communication overlap coefficient corresponding to any audio-visual evaluation segment is the average value of the communication overlap coefficients of each speech segment in the audio-visual evaluation segment;

[0067] Separating the audio at the time corresponding to the audio-visual evaluation segment to obtain a number of audio segments;

[0068] For any speech segment, the communication overlap coefficient of the speech segment is determined according to the number of overlapping speech segments and the proportion of overlapping segments.

[0069] Among them, each speech segment obtained by separation corresponds to only one communicator and the communicator has continuous changes in communication characteristics in the corresponding speech segment. In addition, each speech segment corresponds to a speech duration period. The starting time of a speech duration period is the starting time of the speech segment corresponding to the speech duration period, and the ending time of a speech duration period is the ending time of the speech segment corresponding to the speech duration period. How to separate the speech segments for the audio-visual evaluation segment and confirm the speech duration period of each speech segment is easy to understand for those skilled in the art and will not be elaborated here.

[0070] For any speech segment, the speech segment with which the speech duration period overlaps with that of the speech segment is recorded as the overlapping speech segment. The overlapping segment ratio = the duration of the overlapping part of the speech segment and other speech segments / the duration of the speech duration period of the speech segment. For any speech segment and any overlapping speech segment, the overlapping segment is the overlapping part of the speech duration period corresponding to the speech segment and the speech duration period corresponding to the overlapping speech segment. The overlapping coefficient is the natural logarithm of the product of the number of overlapping speech segments and the overlapping segment ratio.

[0071] Specifically, the paragraph evaluation strategy is determined according to the category of each audio-visual evaluation paragraph;

[0072] For any audio and video evaluation paragraph,

[0073] If the audio-visual evaluation segment is a first-class evaluation segment, a synchronous evaluation coefficient of the audio-visual evaluation segment is determined by a feature matching method;

[0074] If the audio-visual evaluation segment is a second-class evaluation segment, the synchronization evaluation coefficient of the audio-visual evaluation segment is determined by using a timing matching method, and the synchronization evaluation coefficient is determined according to a time difference reference value between each voice segment and its corresponding video segment.

[0075] Among them, if the synchronization evaluation coefficient of the audio-visual evaluation segment is determined by a time matching method, a voice segment is extracted for a audio-visual evaluation segment, and the video segment to be analyzed corresponding to each voice segment is determined according to the voice duration period of each voice segment. For a single video segment to be analyzed, a number of video frames are randomly obtained, and the communication features of the communicators in each video frame are identified. For a single video frame, the communication features of the video frame and the feature similarity of the preset communication features of the characters at the corresponding moment of the video frame in the language segment are detected. The synchronization parameter of the video segment to be analyzed and the corresponding language segment is the average value of the feature similarities corresponding to each video frame. The synchronization evaluation coefficient of the audio-visual evaluation segment is positively correlated with the average value of the synchronization parameters of each video segment to be analyzed and the corresponding language segment;

[0076] The communication features in the present invention are the lip shape features of the communicators. Any character is set with a corresponding preset communication feature. How to set the corresponding preset communication feature for each character and how to detect the feature similarity between the communication features of the communicators and the preset communication features of each character are contents that are easy to understand for those skilled in the art and will not be elaborated here.

[0077] Specifically, when determining the synchronization evaluation coefficient of any audio-visual evaluation segment by using the feature matching method, the phase overlap coefficient of each video segment within the speech duration period corresponding to each speech segment of the audio-visual evaluation segment is detected, and the video segment with a phase overlap coefficient greater than a preset phase overlap coefficient is recorded as the video segment to be evaluated of the corresponding speech segment;

[0078] The feature matching coefficients of each video segment to be evaluated and the corresponding speech segment are detected, and the average value of the feature matching coefficients of each video segment to be evaluated is recorded as the reference feature matching coefficient of the corresponding speech segment;

[0079] The synchronous evaluation coefficient of any audio-visual evaluation segment is positively correlated with the number of feature matching segments, and the feature matching segment is a speech segment whose reference feature matching coefficient is greater than a preset reference feature matching coefficient.

[0080] Among them, for any voice segment, the video corresponding to the voice duration period corresponding to the voice segment is extracted, and the video is divided according to the communication personnel to obtain several video segments, and any video segment corresponds to only one communication personnel. For a single video segment, the stage overlap coefficient of the corresponding voice segment is the sum of the starting difference ratio and the ending difference ratio, the starting difference ratio = the interval time between the starting time of the video segment and the voice segment / the duration of the voice duration period of the voice segment, the ending difference ratio = the interval time between the ending time of the video segment and the voice segment / the duration of the voice duration period of the voice segment; the value of the preset stage overlap coefficient can be determined by the user according to the actual work scenario, for example, the user can set it according to the historical evaluation records, the higher the user's requirements for the lossless effect of video compression, the larger the value of the preset stage overlap coefficient, and a method for determining the value of the preset stage overlap coefficient is provided, and the average value of the stage overlap coefficients of each video segment to be evaluated in the historical evaluation records that meet the user's requirements for the lossless effect of video compression is recorded as the preset stage overlap coefficient;

[0081] The value of the preset reference feature matching coefficient can be determined by the user according to the actual working scenario. For example, the user can set it according to the historical evaluation records. The higher the user's requirement for the lossless effect of video compression, the larger the value of the preset reference feature matching coefficient. A method for determining the value of the preset reference feature matching coefficient is provided, and the average value of the reference feature matching coefficients of each feature matching paragraph in the historical evaluation records that meet the user's requirement for the lossless effect of video compression is recorded as the preset reference feature matching coefficient.

[0082] Specifically, the feature matching coefficient of any video segment to be evaluated and the corresponding speech segment is determined according to the output rate matching degree of the video segment to be evaluated and the corresponding speech segment and the reference matching coefficient;

[0083] The characteristic matching coefficient is positively correlated with the output rate matching degree and the reference matching coefficient respectively.

[0084] Specifically, the output rate matching degree is determined according to the standard matching difference value of the video segment to be evaluated, and the output rate matching degree is negatively correlated with the standard matching difference value;

[0085] The reference matching coefficient is negatively correlated with an average value of the relevant feature interval durations of each video frame of the video segment to be evaluated.

[0086] Among them, the feature matching coefficient of any video segment to be evaluated and the corresponding voice segment is the sum of the output rate matching degree and the reference matching coefficient, the standard matching difference value = output rate matching parameter - standard rate matching parameter, output rate matching parameter = the number of changes in the communication characteristics of the communicators per unit time / the number of characters contained in the voice segment per unit time, and a method for obtaining the value of the standard rate matching parameter is provided, which will meet the user's requirements for the lossless effect of video compression in the historical evaluation record of the voice segment to be evaluated. The average value of the maximum value of the output rate matching parameter in the video segment to be evaluated;

[0087] For any video segment to be evaluated, several video frames are randomly extracted, and the communication features corresponding to each video frame are identified. For a single video frame, the characters in the speech segment whose feature similarity with the communication features corresponding to the video frame is greater than the preset feature similarity are recorded as relevant characters. The relevant feature interval duration is the minimum value of the interval duration between the corresponding moment of the video frame and the corresponding moment of each relevant character in the speech segment. The value of the preset feature similarity can be determined by the user according to the actual working scenario. For example, the user can set it according to the historical evaluation records. The higher the user's requirements for the lossless effect of video compression, the higher the value of the preset feature similarity. A method for determining the value of the preset feature similarity is provided, and the average value of the feature similarities of each relevant character in the historical evaluation records that meet the user's requirements for the lossless effect of video compression is recorded as the preset feature similarity.

[0088] Specifically, under the evaluation completion condition, the interactive effective parameters corresponding to each audio and video evaluation paragraph of the target evaluation video are detected, and the paragraph focus coefficient is determined according to the interactive effective parameters and the synchronization evaluation coefficient of each audio and video evaluation paragraph, and the audio and video evaluation paragraph with a paragraph focus coefficient greater than the preset paragraph focus coefficient is recorded as a key paragraph;

[0089] The interaction validity parameter is determined according to the reference interaction amount and interaction richness of the corresponding synchronous evaluation paragraph.

[0090] Among them, for any audio and video evaluation paragraph, the paragraph focus coefficient is the difference between the product of the interactive effective parameter of the audio and video evaluation paragraph and the corresponding focus influence coefficient and the product of the synchronous evaluation coefficient and the corresponding focus influence coefficient. The values ​​of the interactive effective parameter and the focus influence coefficient corresponding to the synchronous evaluation coefficient can be determined by the user according to the actual work scenario, and a value of the focus influence coefficient corresponding to the interactive effective parameter and the synchronous evaluation coefficient is provided. The value of the focus influence coefficient corresponding to the interactive effective parameter is 0.5, and the value of the focus influence coefficient corresponding to the synchronous evaluation coefficient is 0.5; for any audio and video evaluation paragraph, the reference interaction amount is the number of interactions within the time period corresponding to the audio and video evaluation paragraph, and the interaction richness = the number of effective interactions / reference interaction amount, and the interactive effective parameter is the reference interaction amount and the interaction richness. The natural logarithm of the product of , for any interaction, if the number of repeated words in its interactive content is less than the preset number of repeated words, then the interaction is recorded as a valid interaction, and the number of repeated words is the maximum value of the number of consecutive words repeated with other interactive contents. The value of the preset number of repeated words can be determined by the user according to the actual work scenario. For example, the user can set it according to the historical evaluation records. The higher the user's requirements for the lossless effect of video compression, the smaller the value of the preset number of repeated words. A method for setting the preset number of repeated words is provided, and the average value of the number of repeated words for each valid interaction in the historical evaluation records that meet the user's requirements for the lossless effect of video compression is recorded as the preset number of repeated words. The forms of interaction in the present invention include but are not limited to: barrage and comments, and the interactive content is the text content in the barrage or comments.

[0091] Specifically, under the condition of key analysis completion, the execution compression strategy of the target evaluation video is determined according to the proportion of key paragraphs.

[0092] If the proportion of key paragraphs is greater than the preset proportion of key paragraphs, a dynamic execution strategy is used to compress the target evaluation video;

[0093] If the proportion of key paragraphs is less than or equal to the preset proportion of key paragraphs, a static execution strategy is used to compress the target evaluation video.

[0094] Among them, the key paragraph ratio = the number of key paragraphs in the target evaluation video / the number of audio and video evaluation paragraphs in the target evaluation video. The value of the preset key paragraph ratio can be determined by the user according to the actual work scenario. For example, the user can set it according to the historical evaluation record. The higher the user's requirement for the lossless effect of video compression, the smaller the value of the preset key paragraph ratio. A method for determining the preset key paragraph ratio is provided. The historical evaluation record of the compression of the target evaluation video using a dynamic execution strategy is recorded as a ratio reference record. The minimum value of the key paragraph ratio in the ratio reference record that meets the user's requirement for the lossless effect of video compression is recorded as the preset key paragraph ratio. A value of the preset key paragraph ratio is provided, and the value of the preset key paragraph ratio is 0.3;

[0095] If a static execution strategy is used to compress the target evaluation video, fixed compression parameters are used to complete the compression process of the target evaluation video. The compression parameters are not adjusted during the compression process. Users can set the fixed compression parameters according to the actual working scenario. This is easy for technical personnel in this field to understand and will not be elaborated here.

[0096] Specifically, if a dynamic execution strategy is used to compress the target evaluation video, the video experience parameters of some target evaluation videos that have completed compression in each video detection cycle are periodically detected, and it is determined whether to adjust the compression parameters according to the video experience parameters;

[0097] If the video experience parameter is less than or equal to the preset video experience parameter, determining to adjust the compression parameter of the next video detection cycle of the video detection cycle;

[0098] If the video experience parameter is greater than the preset video experience parameter, it is determined not to adjust the compression parameter of the next video detection cycle of the video detection cycle.

[0099] Among them, for any video detection cycle, the video experience parameter is determined according to the standard peak signal-to-noise ratio difference value and the number of frame skipping of the target evaluation video of the part that completes compression within the video detection cycle, and the video experience parameter=ln(standard peak signal-to-noise ratio difference value / number of frame skipping times), the number of frame skipping times is the absolute value of the difference between the number of frames actually displayed and the number of frames that should be displayed theoretically within the video detection cycle, the standard peak signal-to-noise ratio difference value is the absolute value of the difference between the peak signal-to-noise ratio and the standard peak signal-to-noise ratio, how to detect the peak signal-to-noise ratio and the number of frame skipping times are all easy to understand for those skilled in the art, and will not be elaborated here; the user can set the value of the standard peak signal-to-noise ratio according to the actual working scenario, and provide a standard peak signal-to-noise ratio value, the standard peak signal-to-noise ratio value is 35db;

[0100] In the present invention, a cyclic video detection cycle is applied, and the duration of the video detection cycle can be determined by the user. The higher the user's requirements for the lossless effect of video compression, the shorter the duration of the video detection cycle. A duration of a video detection cycle is provided, and the video detection cycle is 2 minutes. At the end of each video detection cycle, the video experience parameters of the compressed part of the video completed in the video detection cycle are detected, and it is determined whether to adjust the compression parameters for the next video detection cycle according to the video experience parameters;

[0101] The value of the preset video experience parameter can be determined by the user according to the actual working scenario. For example, the user can set it according to the historical evaluation record. The higher the user's requirement for the lossless effect of video compression, the smaller the preset video experience parameter. A method for taking the value of the preset video experience parameter is provided. The historical evaluation record corresponding to the video detection period without compression parameter adjustment is recorded as a parameter reference record. The minimum value of the video experience parameter of each video detection period in the parameter reference record that meets the user's requirement for the lossless effect of video compression is recorded as the preset video experience parameter. The compression parameters that can be adjusted in the present invention include but are not limited to: frame rate and bit rate. A method for adjusting the compression parameter is provided. If the video experience parameter is less than or equal to the preset video experience parameter, it is determined to increase the frame rate and bit rate of the next video detection period. The increase value of the frame rate is negatively correlated with the video experience parameter, and the increase value of the bit rate is negatively correlated with the video experience parameter.

[0102] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0103] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A digital evaluation method for film and television videos, characterized in that: include: Extract audio and video evaluation segments from the target evaluation video, and detect the communication complexity coefficient and reference communication overlap coefficient of each audio and video evaluation segment; Determine the category of the audio-visual evaluation paragraph according to the communication complexity coefficient and the reference communication coincidence coefficient, and determine the corresponding paragraph evaluation strategy according to the category of each audio-visual evaluation paragraph; If an audio-visual evaluation paragraph is a type of evaluation paragraph, a feature matching method is used to determine a synchronization evaluation coefficient, and the synchronization evaluation coefficient of the corresponding audio-visual evaluation paragraph is determined according to the number of feature matching paragraphs; If the audio and video evaluation segment is a second-class evaluation segment, the synchronization evaluation coefficient is determined by using a time sequence matching method, and the synchronization evaluation coefficient is determined according to a time difference reference value between each audio segment and its corresponding video segment; Under the evaluation completion condition, the segment focus coefficient is determined according to the interactive effective parameters and synchronous evaluation coefficients of each audio and video evaluation segment. The evaluation completion condition is that the synchronous evaluation coefficients of each audio and video evaluation segment in the target evaluation video are all determined and completed; Under the condition of key analysis completion, the execution compression strategy of the target evaluation video is determined according to the proportion of key paragraphs. The execution compression strategy is to use a dynamic execution strategy or a static execution strategy to compress the target evaluation video. The condition of key analysis completion is that all key paragraphs in the target evaluation video are determined to be completed; The communication complexity coefficient is determined according to the number of communicators in the audio-visual evaluation section and the reference communication frequency. The reference communication frequency is the average of the communication frequencies of the communicators. For a single communicator, the communication frequency = the number of changes in the communication characteristics of the communicator in the audio-visual evaluation section / the duration of the communication of the communicator in the audio-visual evaluation section. The duration of the communication is the duration of the communication characteristics of the communicator in the audio-visual evaluation section. For any communicator, if the communication characteristics collected by two adjacent video frames are inconsistent, it is recorded as a change in the communication characteristics. For any audio-visual evaluation segment, the communication complexity coefficient is the natural logarithm of the product of the number of communicators in the audio-visual evaluation segment and the reference communication frequency of the audio-visual evaluation segment; The reference communication overlap coefficient corresponding to any audio-visual evaluation segment is the average value of the communication overlap coefficients of each speech segment in the audio-visual evaluation segment; The interaction validity parameter is determined according to the reference interaction amount and interaction richness of the corresponding synchronous evaluation paragraph. For any audio-visual evaluation paragraph, the reference interaction amount is the number of interactions in the time period corresponding to the audio-visual evaluation paragraph, and the interaction richness = the number of effective interactions / reference interaction amount. The interaction validity parameter is the natural logarithm of the product of the reference interaction amount and the interaction richness; If the synchronization evaluation coefficient of the audio-visual evaluation segment is determined by using a time matching method, a speech segment is extracted for an audio-visual evaluation segment, and the video segment to be analyzed corresponding to each speech segment is determined according to the speech duration period of each speech segment. For a single video segment to be analyzed, a number of video frames are randomly obtained, and the communication features of the communicators in each video frame are identified. For a single video frame, the communication features of the video frame and the feature similarity of the preset communication features of the characters at the corresponding moment of the video frame in the language segment are detected. The synchronization parameter of the video segment to be analyzed and the corresponding language segment is the average value of the feature similarities corresponding to each video frame. The synchronization evaluation coefficient of the audio-visual evaluation segment is positively correlated with the average value of the synchronization parameters of each video segment to be analyzed and the corresponding language segment; When determining the synchronization evaluation coefficient of any audio-visual evaluation segment by using a feature matching method, the phase overlap coefficient of each video segment within the speech duration period corresponding to each speech segment of the audio-visual evaluation segment is detected, and the video segment whose phase overlap coefficient is greater than the preset phase overlap coefficient is recorded as the video segment to be evaluated of the corresponding speech segment; Detect the feature matching coefficients of each video segment to be evaluated and the corresponding speech segment, and determine the reference feature matching coefficient of the corresponding speech segment according to the feature matching coefficients of each video segment to be evaluated; The synchronous evaluation coefficient of any audio-visual evaluation segment is positively correlated with the number of feature matching segments, wherein the feature matching segment is a speech segment whose reference feature matching coefficient is greater than a preset reference feature matching coefficient, and the reference feature matching coefficient is an average value of the feature matching coefficients of each video segment to be evaluated; For a single video segment, the phase overlap coefficient of the corresponding speech segment is the sum of the starting difference ratio and the ending difference ratio, the starting difference ratio = the interval time between the starting moment of the video segment and the speech segment / the duration of the speech duration period of the speech segment, the ending difference ratio = the interval time between the ending moment of the video segment and the speech segment / the duration of the speech duration period of the speech segment, the feature matching coefficient of any video segment to be evaluated and the corresponding speech segment is the sum of the output rate matching degree and the reference matching coefficient, the output rate matching degree is negatively correlated with the standard matching difference value, the reference matching coefficient is negatively correlated with the average value of the relevant feature interval duration of each video frame of the video segment to be evaluated, the standard matching difference value = output rate matching parameter-standard rate matching parameter, output rate matching parameter = the number of changes in the communication features of the communicators per unit time / the number of characters contained in the speech segment per unit time, and the relevant feature interval duration is the minimum value of the interval duration between the corresponding moment of the video frame and the corresponding moment of each relevant character in the speech segment.

2. The digital evaluation method of film and television videos according to claim 1, characterized in that: Determine the category of the audio-visual evaluation paragraph according to the communication complexity coefficient and the reference communication coincidence coefficient; If the communication complexity coefficient of an audio-visual evaluation paragraph is greater than the preset communication complexity coefficient or the reference communication coincidence coefficient is greater than the preset communication coincidence coefficient, the audio-visual evaluation paragraph is determined to be a first-class evaluation paragraph; If the communication complexity coefficient of an audio-visual evaluation segment is less than or equal to the communication complexity coefficient and the reference communication overlap coefficient is less than or equal to the preset communication overlap coefficient, the audio-visual evaluation segment is determined to be a second-class evaluation segment.

3. The digital evaluation method of film and television video according to claim 2, characterized in that: Separating the audio at the time corresponding to the audio-visual evaluation segment to obtain a number of audio segments; For any speech segment, the communication overlap coefficient of the speech segment is determined according to the number of overlapping speech segments and the proportion of overlapping segments; For any speech segment, the speech segment that overlaps with the speech duration of the speech segment is recorded as the overlapping speech segment. The overlapping segment ratio = the duration of the overlapping part of the speech segment and other speech segments / the duration of the speech duration of the speech segment. For any speech segment and any overlapping speech segment, the overlapping segment is the overlapping part of the speech duration corresponding to the speech segment and the speech duration corresponding to the overlapping speech segment. The overlap coefficient is the natural logarithm of the product of the number of overlapping speech segments and the overlapping segment ratio.

4. The digital evaluation method of film and television video according to claim 3 is characterized in that: Under the evaluation completion condition, the interactive effective parameters corresponding to each audio and video evaluation paragraph of the target evaluation video are detected, and the paragraph focus coefficient is determined according to the interactive effective parameters and the synchronization evaluation coefficient of each audio and video evaluation paragraph, and the audio and video evaluation paragraph with a paragraph focus coefficient greater than the preset paragraph focus coefficient is recorded as a key paragraph.

5. The digital evaluation method of film and television video according to claim 4, characterized in that: Under the condition of key analysis completion, the execution compression strategy of the target evaluation video is determined according to the proportion of key paragraphs. If the proportion of key paragraphs is greater than the preset proportion of key paragraphs, a dynamic execution strategy is used to compress the target evaluation video; If the proportion of key paragraphs is less than or equal to the preset proportion of key paragraphs, a static execution strategy is used to compress the target evaluation video.

6. The digital evaluation method of film and television video according to claim 5, characterized in that: If a dynamic execution strategy is used to compress the target evaluation video, the video experience parameters of some target evaluation videos that have completed compression in each video detection cycle are periodically detected, and it is determined whether to adjust the compression parameters according to the video experience parameters; If the video experience parameter is less than or equal to the preset video experience parameter, determining to adjust the compression parameter of the next video detection cycle of the video detection cycle; If the video experience parameter is greater than the preset video experience parameter, it is determined not to adjust the compression parameter of the next video detection cycle of the video detection cycle; For any video detection cycle, the video experience parameter is determined according to the standard peak signal-to-noise ratio difference value and the number of frame skipping of the target evaluation video of the part that completes compression within the video detection cycle. The video experience parameter=ln(standard peak signal-to-noise ratio difference value / number of frame skipping times), the number of frame skipping times is the absolute value of the difference between the number of frames actually displayed within the video detection cycle and the number of frames that should be displayed theoretically, and the standard peak signal-to-noise ratio difference value is the absolute value of the difference between the peak signal-to-noise ratio and the standard peak signal-to-noise ratio.

Images