CN112804586B - Method, device and equipment for acquiring video clip - Google Patents
Method, device and equipment for acquiring video clip Download PDFInfo
- Publication number
- CN112804586B CN112804586B CN202110392468.XA CN202110392468A CN112804586B CN 112804586 B CN112804586 B CN 112804586B CN 202110392468 A CN202110392468 A CN 202110392468A CN 112804586 B CN112804586 B CN 112804586B
- Authority
- CN
- China
- Prior art keywords
- frame image
- test question
- temporary
- step length
- determining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
- H04N21/44—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream, rendering scenes according to MPEG-4 scene graphs
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
- H04N21/44—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream, rendering scenes according to MPEG-4 scene graphs
- H04N21/44008—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream, rendering scenes according to MPEG-4 scene graphs involving operations for analysing video streams, e.g. detecting features or characteristics in the video stream
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
- H04N21/83—Generation or processing of protective or descriptive data associated with content; Content structuring
- H04N21/845—Structuring of content, e.g. decomposing content into time segments
- H04N21/8455—Structuring of content, e.g. decomposing content into time segments involving pointers to the content, e.g. pointers to the I-frames of the video stream
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
- H04N21/83—Generation or processing of protective or descriptive data associated with content; Content structuring
- H04N21/845—Structuring of content, e.g. decomposing content into time segments
- H04N21/8456—Structuring of content, e.g. decomposing content into time segments by decomposing the content in the time domain, e.g. in time segments
Abstract
The disclosure relates to a method, a device and equipment for acquiring a video clip, wherein the method comprises the following steps: acquiring an original video; repeatedly executing the following positioning operation until the positioning operation is executed to the end of the original video, and stopping: identifying a first frame image containing test questions in an original video based on a preset step length adjustment strategy; determining a starting frame image containing the test question based on the initial step length and the position of the first frame image; identifying a second frame image which does not contain the test question after the initial frame image based on a step length adjusting strategy, and determining an ending frame image containing the test question based on the initial step length or the position of the initial frame image and the position of the second frame image; thereafter, with respect to the frame image subsequent to the termination frame image, the first frame image at the time of next performing the positioning operation is newly determined. And when the positioning operation is stopped, acquiring all target video clips containing the test questions in the original video. The method and the device can improve the acquisition efficiency and accuracy of the target video clip.
Description
Technical Field
The present disclosure relates to the field of video processing technologies, and in particular, to a method, an apparatus, and a device for acquiring a video clip.
Background
With the popularization of video courses, it is becoming an important requirement to intercept a topic video clip in a video course. At present, picture information in a video can be analyzed frame by frame or with a certain step length, and then editing is performed according to the analyzed picture information; however, the frame-by-frame analysis method is time-consuming and inefficient, and the method of analyzing with a certain step size has a problem of low accuracy in locating the clipped video segment.
Disclosure of Invention
In order to solve the technical problem or at least partially solve the technical problem, the present disclosure provides a method, an apparatus, and a device for acquiring a video clip, which can improve the acquisition efficiency and accuracy of a target video clip.
The present disclosure provides a method for acquiring a video clip, including:
acquiring an original video; repeatedly executing the following positioning operation until the positioning operation is executed to the end of the original video, and stopping:
identifying a first frame image containing test questions in the original video based on a preset step length adjustment strategy; the step length adjusting strategy is a strategy for adjusting the advancing step length based on the result of whether the frame image contains the test question or not; determining a starting frame image of a video segment containing the test question based on the initial step length and the position of the first frame image; identifying a second frame image which does not contain the test question after the initial frame image based on the step length adjusting strategy, and determining an ending frame image of the video segment containing the test question based on the initial step length or the position of the initial frame image and the position of the second frame image; returning to the step of identifying the first frame image containing the test question in the original video based on a preset step length adjusting strategy aiming at the frame image behind the termination frame image in the original video;
when the positioning operation is stopped, acquiring all target video clips in the original video; wherein the target video segment is a video segment containing a test question determined based on the start frame image and the end frame image.
Optionally, the step size adjustment policy is:
wherein, a1Adjusting the step length adjustment coefficient in the strategy for the first step length, and a1＜0，a2A step size adjustment coefficient in the second step size adjustment strategy, and2＞0，stepistep for the current step sizei+1Step for the next step after the current step is adjustediAnd stepi+1Are constrained within a preset step range.
Optionally, the method includes:
identifying whether the current frame image to be identified contains test questions; if the frame image to be identified contains the test question, determining the next target step length based on the current step length and the first step length adjusting strategy; if the frame image to be identified does not contain the test question, determining the next target step length based on the current step length and a second step length adjusting strategy; and determining the frame image which is separated from the frame image to be identified by the target step length as a new frame image to be identified, and returning to the step of identifying whether the current frame image to be identified contains the test question or not.
Optionally, the identifying whether the current frame image to be identified contains the test question includes:
carrying out feature extraction on a frame image to be identified through a feature extraction network to obtain a feature image; recognizing text information of the characteristic image according to the optical character recognition COR, and converting the text information into word vectors; judging whether the word vector is a test question or not according to a preset question discriminator; and when the word vector is a test question, determining that the frame image to be identified contains the test question.
Optionally, the identifying a first frame image containing a test question in the original video based on a preset step size adjustment strategy includes: repeatedly executing the following identification operation until the temporary frame image is identified to contain the test question, and determining a first frame image containing the test question:
identifying whether a temporary frame image in the original video contains a test question or not; when the positioning operation is executed for the first time, the temporary frame image is a preset frame image of the original video; when the positioning operation is not executed for the first time, the temporary frame image is a frame image which is separated from an ending frame image determined when the positioning operation is executed for the last time by an initial step length; if the temporary frame image does not contain the test question, determining the forward step length of the video forward based on a preset step length adjustment strategy, and determining the frame image which is separated from the temporary frame image by the forward step length as a new temporary frame image; and if the temporary frame image contains the test question, determining the temporary frame image as a first frame image.
Optionally, the determining a starting frame image of a video segment containing a test question based on the initial step size and the position of the first frame image includes:
based on the step length adjustment strategy, identifying a third frame image which is before the first frame image and is separated from the first frame image by the nearest distance and does not contain the test question, and taking the position of the third frame image as a first temporary starting point of a test question starting point searching range; taking the position of the first frame image as a first temporary end point of a test question starting point searching range; based on the first temporary starting point and the first temporary ending point, repeatedly performing a first adjustment operation as follows:
determining a first temporary intermediate point between the first temporary start point and the first temporary end point, and determining a first candidate image of a frame next to the first temporary intermediate point; respectively identifying whether the frame image corresponding to the first temporary intermediate point and the first candidate image contain test questions; if the frame image corresponding to the first temporary intermediate point and the first candidate image both contain test questions, adjusting the first temporary intermediate point to be a new first temporary end point; if the frame image corresponding to the first temporary intermediate point and the first candidate image do not contain the test question, adjusting the position of the first candidate image to be a new first temporary starting point;
and repeatedly executing the first adjustment operation until the frame image corresponding to the first temporary intermediate point does not contain the test question and the first candidate image contains the test question, stopping the first adjustment operation, and determining the first candidate image corresponding to the first adjustment operation when the first adjustment operation is stopped as the initial frame image of the video segment containing the test question.
Optionally, the determining, based on the initial step size or the position of the start frame image and the position of the second frame image, an end frame image of the video segment containing the test question includes:
based on the step length adjustment strategy, identifying a fourth frame image which is before the second frame image and is separated from the second frame image by the nearest test question, and taking the position of the fourth frame image as a second temporary starting point of a test question end point searching range; taking the position of the second frame image as a second temporary end point of the test question end point searching range; based on the second temporary starting point and the second temporary ending point, repeatedly performing a second adjustment operation as follows:
determining a second temporary intermediate point between the second temporary start point and the second temporary end point, and determining a second candidate image of a frame next to the second temporary intermediate point; respectively identifying whether the frame image corresponding to the second temporary intermediate point and the second candidate image contain test questions; if the frame image corresponding to the second temporary intermediate point and the second candidate image both contain the test question, adjusting the position of the second candidate image to be a new second temporary starting point; if the frame image corresponding to the second temporary intermediate point and the second candidate image do not contain the test question, adjusting the second temporary intermediate point to be a new second temporary end point;
and repeatedly executing the second adjustment operation until the frame image corresponding to the second temporary intermediate point contains the test question and the second candidate image does not contain the test question, stopping the second adjustment operation, determining the corresponding second temporary intermediate point when the second adjustment operation is stopped as a target end point, and determining the frame image corresponding to the target end point as an end frame image of the video clip containing the test question.
Optionally, the method further includes: after a target video segment containing test questions is determined, generating a title of the target video segment according to a preset title generator, and adding the title to the target video segment.
The present disclosure provides an apparatus for acquiring a video clip, including:
the original video acquisition module is used for acquiring an original video;
a positioning operation repeated execution module, configured to repeatedly execute the following positioning operations until the positioning operations are executed until the end of the original video:
the identification unit is used for identifying a first frame image containing the test question in the original video based on a preset step length adjustment strategy; the step length adjusting strategy is a strategy for adjusting the advancing step length based on the result of whether the frame image contains the test question or not;
the starting frame positioning unit is used for determining a starting frame image of a video segment containing the test question based on the initial step length and the position of the first frame image;
the termination frame positioning unit is used for identifying a second frame image which does not contain the test question after the initial frame image based on the step length adjusting strategy, and determining a termination frame image of a video segment containing the test question based on the initial step length or the position of the initial frame image and the position of the second frame image; and returning to the step executed by the identification unit for a frame image subsequent to the termination frame image of the original video;
a target video clip obtaining module, configured to obtain all target video clips in the original video when the positioning operation is stopped; wherein the target video segment is a video segment containing a test question determined based on the start frame image and the end frame image.
The present disclosure provides an electronic device, the electronic device including: a processor and a storage device; the storage means has stored thereon a computer program which, when executed by the processor, performs the method described above.
Compared with the prior art, the technical scheme provided by the embodiment of the disclosure has the following advantages:
the method comprises the step of repeatedly executing positioning operation until the positioning operation is executed to the end of an original video, so that all target video clips containing test questions in the original video are obtained. The positioning operation is as follows: the method comprises the steps that firstly, a first frame image containing test questions in an original video is identified based on a preset step length adjusting strategy; secondly, determining a starting frame image of the video segment containing the test question based on the initial step length and the position of the first frame image; identifying a second frame image which does not contain the test question after the initial frame image based on a step length adjusting strategy, and determining an ending frame image of a video clip containing the test question based on the initial step length or the position of the initial frame image and the position of the second frame image; thereafter, for the frame image after the terminating frame image of the original video, the first step is returned to determine the new first frame image when the positioning operation is performed next time.
In the above-mentioned solution provided in this embodiment, on one hand, in the process of repeatedly performing the following positioning operation, the step length adjustment policy can flexibly adjust the advance step length according to whether the frame image contains the test question, so as to effectively improve the efficiency and accuracy of acquiring the first frame image containing the test question and the second frame image not containing the test question. On the other hand, based on the position and the initial step length of the first frame image, the initial frame image can be more accurately positioned, the initial frame image is utilized to further position the termination frame image containing the test question, and the positioning accuracy of the termination frame image can be improved; therefore, the target video segment determined based on the starting frame image and the ending frame image can contain the test questions and has high-accuracy positioning. Therefore, the scheme improves the acquisition efficiency and accuracy of the target video clip.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure.
In order to more clearly illustrate the embodiments or technical solutions in the prior art of the present disclosure, the drawings used in the description of the embodiments or prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.
Fig. 1 is a flowchart of a method for acquiring a video segment according to an embodiment of the present disclosure;
fig. 2 is a schematic diagram of a search range of a test question starting point provided in the embodiment of the present disclosure;
fig. 3 is a flowchart of another method for acquiring a video segment according to an embodiment of the present disclosure;
fig. 4 is a block diagram of a structure of an apparatus for acquiring a video segment according to an embodiment of the present disclosure.
Detailed Description
In order that the above objects, features and advantages of the present disclosure may be more clearly understood, aspects of the present disclosure will be further described below. It should be noted that the embodiments and features of the embodiments of the present disclosure may be combined with each other without conflict.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present disclosure, but the present disclosure may be practiced in other ways than those described herein; it is to be understood that the embodiments disclosed in the specification are only a few embodiments of the present disclosure, and not all embodiments.
In view of the inefficiency and the low accuracy of the conventional method for acquiring a video clip containing a test question, embodiments of the present disclosure provide a method, an apparatus, and a device for acquiring a video clip, and for understanding, the embodiments of the present disclosure are described in detail below.
The first embodiment is as follows:
referring to fig. 1, a flow chart of a method for acquiring a video segment is provided, where the method may include the following steps:
step S102, acquiring an original video. In this embodiment, the original video may be a video shot by the video image capturing device, or may be a video downloaded through a network, stored locally, or uploaded manually, and the original video is a video containing test question content.
For the original video, the following positioning operations shown in steps S104 to S108 may be repeatedly performed until the positioning operation is performed to the end of the original video, and then stopped:
and step S104, identifying a first frame image containing the test question in the original video based on a preset step length adjustment strategy.
The step length adjustment strategy is a strategy for adjusting the advancing step length based on the result of whether the frame image contains the test question or not, and is shown in the following formula:
wherein, a1Adjusting the step size of the strategy for the first step size, and1＜0，a2for the step size adjustment coefficient in the second step size adjustment strategy, and a2Is greater than 0, and the step length adjustment coefficient a is used for simplifying the step length adjustment strategy1And a2Can be mutually opposite numbers, e.g. a1=-0.2，a2=0.2。stepiStep for the current step sizei+1Step for the next step after the current step is adjustediAnd stepi+1Are constrained within a predetermined step size range, for example, (0.2 × step)1）≤stepi≤（5*step1），step1For example, set to 30s or 30 frames for the initial step size.
The above adjustment strategy represents: if the current frame image to be identified contains the test question, the probability that the test question is contained in a section of continuous frame image taking the current frame image to be identified as a starting point is higher, so that the forward step length of the next frame image to be searched can be reduced compared with the current step length, and the frame image containing the test question is prevented from being missed; in this case, the next target step size may be determined based on the current step size and the first step size adjustment strategy. Correspondingly, if the frame image to be identified does not contain the test question, the probability of containing the test question in a section of continuous frame image taking the current frame image to be identified as a starting point is low, so that the forward step length of the next frame image to be searched is increased, and the frame image not containing the test question is quickly skipped; in this case, the next target step size is determined based on the current step size and the second step size adjustment strategy.
In this embodiment, based on the step size adjustment strategy, the sequence of frame images in the original video from front to back is advanced at least once to obtain the first frame image containing the test question.
Step S106, based on the initial step length and the position of the first frame image, determining a starting frame image of the video segment containing the test question.
Since the first frame image containing the test question is determined based on the step adjustment strategy, it is highly likely that the frame image containing the test question also exists in the frame images that are spanned before the first frame image. It can be determined that the real starting point of the test question should be located in the forward video range including the first frame image. In this case, the embodiment may determine the search range of the starting point of the test question based on the position and the initial step length of the first frame image, and determine the starting frame image containing the test question ordered most forward in the search range of the starting point of the test question, so as to improve the positioning accuracy of the starting frame image.
Step S108, identifying a second frame image which does not contain the test question after the initial frame image based on the step length adjustment strategy, and determining an ending frame image of the video segment containing the test question based on the initial step length or the position of the initial frame image and the position of the second frame image.
In this embodiment, in order to improve the recognition efficiency, the position of the start frame image may be used as a starting point, and the first second frame image without the test question after the start frame image is obtained by advancing at least once along the sequence of the frame images in the original video from front to back based on the step size adjustment strategy.
Similar to the above consideration of determining the search range of the starting point of the test question, the present embodiment can determine that the real end point of the test question should be located in the previous video range including the second frame image. In this case, the test question end point search range may be determined based on the initial step length or the position of the start frame image and the position of the second frame image, and the end frame image including the test question may be determined within the test question end point search range, thereby improving the positioning accuracy of the end frame image.
For the frame image after the terminating frame image in the original video, the process returns to the step S104, and determines a new first frame image when the positioning operation is performed next time. Then, step S106 and step S108 are executed again based on the determined new first frame image, so as to determine the start frame image and the end frame image containing the test question during the new round of positioning operation. And repeating the positioning operation until the time or the frame number corresponding to the determined first frame image exceeds the end position of the original video, determining that the positioning operation is executed until the original video is ended, and stopping the positioning operation.
Step S110, when the positioning operation is stopped, all target video clips in the original video are obtained; the target video segment is a video segment containing the test question determined based on the start frame image and the end frame image, that is, the target video segment is a continuous video segment with the start frame image as a starting point and the end frame image as an end point.
In addition, when the positioning operation is stopped, the corresponding end frame image may not be the end frame image of the original video, and the video segment from the end frame image to the end frame image of the original video is usually short, so that the target video segment containing the test question in the segment of video segment can be extracted in a conventional identification manner, and a description thereof will not be provided.
According to the method for acquiring the video clips, the positioning operation is repeatedly executed until the positioning operation is executed to the end of the original video, so that all target video clips containing test questions in the original video are acquired; the positioning operation is as follows: the method comprises the steps that firstly, a first frame image containing test questions in an original video is identified based on a preset step length adjusting strategy; secondly, determining a starting frame image containing the test question based on the initial step length and the position of the first frame image; identifying a second frame image which does not contain the test question after the initial frame image based on a step length adjusting strategy, and determining an ending frame image of a video clip containing the test question based on the initial step length or the position of the initial frame image and the position of the second frame image; thereafter, for the frame image after the terminating frame image of the original video, the first step is returned to determine the new first frame image when the positioning operation is performed next time.
In the process of repeatedly performing the positioning operation, on one hand, the method for acquiring the video clip provided by this embodiment can flexibly adjust the forward step length according to whether the frame image contains the test question or not by using the step length adjustment strategy, thereby effectively improving the efficiency and accuracy of acquiring the first frame image containing the test question and the second frame image not containing the test question. On the other hand, based on the position and the initial step length of the first frame image, the initial frame image can be more accurately positioned, the initial frame image is utilized to further position the termination frame image containing the test question, and the positioning accuracy of the termination frame image can be improved; therefore, the target video segment determined based on the starting frame image and the ending frame image can contain the test questions and has high-accuracy positioning. Therefore, the scheme improves the acquisition efficiency and accuracy of the target video clip.
In one embodiment, a method for applying a step size adjustment policy is provided, as shown below:
step one, identifying whether a current frame image to be identified contains a test question. The frame image to be identified may be the frame image involved in the process of identifying the first frame image containing the test question and identifying the second frame image not containing the test question in the foregoing embodiment.
In a specific implementation manner, firstly, feature extraction is performed on a frame image to be identified through a feature extraction network (such as resnet) to obtain a feature image; then, recognizing text information of the characteristic image according to OCR (Optical Character Recognition), and converting the text information into word vectors; and finally, judging whether the word vector is a test question or not according to a preset question discriminator, wherein after the word vector is input to the question discriminator, the question discriminator outputs 1 or 0,1 represents that the input word vector is a test question, and 0 represents that the input word vector is not a test question.
Step two, if the frame image to be identified contains the test question, determining the next target step length based on the current step length and the first step length adjustment strategy; and if the frame image to be identified does not contain the test question, determining the next target step length based on the current step length and the second step length adjusting strategy.
In this embodiment, when a word vector is a test question, it is determined that a frame image to be recognized contains the test question; in this case, if it is required to continue to find the next frame image containing or not containing the test question along the sequence of frame images from front to back (or from back to front) in the original video, the step can be based on the current step sizeiAnd the first step length adjustment strategy determines the next target step lengthi+1. It can be understood that the current step size stepiBased on the last stepi-1And the first step length adjustment strategy or the second step length adjustment strategy, and in the two step length adjustment strategies, the initial step length1At a preset value, such as 30 frames。
Referring to the condition that the frame image to be recognized contains the test question, if the frame image to be recognized does not contain the test question, the step is based on the current step lengthiAnd the second step length adjustment strategy determines the next target step lengthi+1And the description is not specifically provided.
And step three, determining the frame image which is separated from the frame image to be identified by the target step length as a new frame image to be identified, and returning to the step one. And by analogy, continuously adjusting the advancing step length according to the step length adjusting strategy, and searching the next frame image according to the adjusted advancing step length.
According to the step size adjustment strategy described above, the method for identifying the first frame image containing the test question in the original video includes repeatedly performing the following identification operations:
and identifying whether the temporary frame image in the original video contains the test question or not. When the positioning operation is executed for the first time, the temporary frame image is a preset frame image of the original video; the preset frame image is such as: the method comprises the steps of obtaining a first frame image of an original video, a specified Nth frame image and a frame image which is separated from the first frame image by an initial step length. When the positioning operation is not performed for the first time, the temporary frame image is a frame image separated from the termination frame image determined when the positioning operation was performed for the last time by an initial step.
If the temporary frame image contains the test question, the temporary frame image can be directly determined as the first frame image.
And if the temporary frame image does not contain the test question, determining the forward step length of the video forward based on a preset step length adjustment strategy, and determining the frame image which is separated from the temporary frame image by the forward step length as a new temporary frame image.
And re-executing the identification operation aiming at the new temporary frame image until the temporary frame image is identified to contain the test questions, and determining the first frame image containing the test questions.
For example, in one possible scenario, when the positioning operation is first performed, the temporary frame image is a frame image (denoted as P1) separated from the first frame image by an initial step size (30 frames) in the original video if it is identified that the temporary frame image P1 does not contain the temporary frame imageTest question, determining the forward step length of the video forward to be step based on the second step length adjusting strategy2=step1*(1+a2) =30 = 1+0.2) =36 frames, and a frame image spaced from the provisional frame image P1 by a forward step size of 36 frames is determined as a new provisional frame image (denoted as P2). If it is recognized that the temporary frame image P2 still contains no test question, step is based on the current step size2And the second step length adjustment strategy adjusts the advancing step length of the video advancing to be step3=step2*(1+a2) =36 × (1+0.2) =43 frames, and a frame image which is spaced from the provisional frame image P2 by a forward step of 43 frames is determined as a new provisional frame image, and so on, and the above-described identification operation is repeatedly performed until the provisional frame image is identified to contain a test question, and the first frame image containing the test question is determined.
In one embodiment, the manner of determining the starting frame image of the video segment containing the test question based on the initial step size and the position of the first frame image can be as follows.
And step I, taking the position which is in front of the first frame image and is separated from the first frame image by an initial step length as a first temporary starting point of the test question starting point searching range. Or, based on the step adjustment strategy, identifying a third frame image which is before the first frame image and is closest to the first frame image and does not contain the test question, and taking the position of the third frame image as a first temporary starting point of the test question starting point searching range.
The position of the first frame picture may be denoted i, and in one possible example the first temporary starting point is before the first frame picture and separated from the first frame picture by an initial step size1Is denoted as s1_ s = i-step1。
In another possible example, the first temporary starting point is based on a step adjustment policy, and a third frame image that does not contain the test question and is before and closest to the first frame image is identified. Specifically, the position i of the first frame image is taken as a starting point, and the step length is adjusted to step according to a first step length adjustment strategy2=step1(1-0.2) =24 frames; finding the frame image of the first frame image which is separated by 24 frames along the sequence of the frame images in the original video from back to frontThe position of the frame image is i-step2. If i-step2If the frame image corresponding to the position contains the test question, searching the next frame image according to the mode, wherein the position of the next frame image is i-step2-step3. And so on until finding the third frame image containing no test question, the position of the third frame image is expressed as s1_1= i-step2-step3。
It should be noted that if the first temporary start point found based on the first frame image exceeds the start point (denoted as 0) of the original video, the first temporary start point is determined as the start point of the original video, i.e., s1_ s = 0.
And II, taking the position of the first frame image as a first temporary end point of the test question starting point searching range. The first temporary end point may be denoted as s1_ e = i.
Referring to fig. 2, assuming that the current frame image is a frame image at the 100 th position in the original video and the preset step size is 30s, the first temporary start point s1_ s of the test question starting point finding range is the 70 th position of the original video, and the first temporary end point s1_ e of the test question starting point finding range is the 100 th position.
Step III, based on the first temporary start point and the first temporary end point, repeatedly performing a first adjustment operation as shown in (1) and (2) below:
(1) a first temporary intermediate point between the first temporary start point and the first temporary end point is determined, and a first candidate image of a frame next to the first temporary intermediate point is determined. Specifically, the first temporary intermediate point may be represented as s1_ c = (s1_ s + s1_ e)/2, and the position of the first candidate image may be represented as s1_{ c +1 }. In fig. 2, the first temporary intermediate point s1_ c is the 85 th position, and the first candidate image s1_ { c +1} refers to the next frame image adjacent to the corresponding frame image at 85 th.
(2) Respectively identifying whether the frame image and the first candidate image corresponding to the first temporary intermediate point contain test questions; referring to table 1, if the frame image corresponding to the first temporary intermediate point s1_ c and the first candidate image s1_ { c +1} both contain a test question, it indicates that there may be a test question in the frame image before the first temporary intermediate point s1_ c, that is, the real starting point of the test question should be located before the first temporary intermediate point s1_ c. In this case, the first temporary intermediate point s1_ c is adjusted to a new first temporary end point s1_ e, and the first temporary start point s1_ s remains unchanged; it is understood that, after adjusting the new first temporary end point s1_ e, the first adjustment operation will be re-executed by returning to step (1) based on the first temporary start point s1_ s and the new first temporary end point s1_ e. If no test question is contained in the frame image corresponding to the first temporary intermediate point s1_ c and the first candidate image s1_ { c +1}, the true start point indicating the test question should be located after the first candidate image s1_ { c +1 }. In this case, the position s1_ { c +1} of the first candidate image is adjusted to the new first temporary start point s1_ s, and the first temporary end point s1_ e remains unchanged. It is understood that, after the new first temporary start point s1_ s is adjusted, the first adjustment operation will be re-executed by returning to step (1) based on the new first temporary start point s1_ s and the first temporary end point s1_ e.
TABLE 1
| s1_c | s1_{c+1} | Adjusting the state |
| Including examination questions | Including examination questions | New s1_ e = = s1_ c |
| Do not contain examination questions | Do not contain examination questions | New s1_ s = = s1_ { c +1} |
| Does not compriseExamination questions | Including examination questions | Stopping the first adjustment operation |
Step IV, the first adjustment operation is repeatedly executed until the frame image corresponding to the first temporary intermediate point s1_ c does not contain any test question and the first candidate image s1_ { c +1} contains a test question, and the first adjustment operation is stopped. The frame image corresponding to the first temporary intermediate point s1_ c does not include a test question and the first candidate image s1_ { c +1} includes a test question, indicating that the first candidate image s1_ { c +1} is the frame image including a test question located closest to the first frame image that is found before the first frame image at the 100 s-th position, and is consecutive to the first frame image, so that the first adjustment operation can be stopped, and the corresponding first candidate image at the time when the first adjustment operation is stopped can be determined as the starting frame image of the video segment including the test question.
In the determining method of the start frame image provided in this embodiment, the search range of the test question start point is gradually reduced by continuously adjusting the first temporary start point, the first temporary end point, and the first temporary intermediate point, so that the start frame image containing the test question is accurately positioned, and the positioning accuracy of the test question start point is improved.
After the initial frame image is obtained, the advance step length is adjusted at least once according to the step length adjustment strategy, and the first second frame image which does not contain the test question behind the initial frame image is searched for according to the adjusted advance step length.
In one embodiment, based on the initial step size or the position of the start frame image and the position of the second frame image, the manner of determining the ending frame image of the video segment containing the test question can be referred to as follows:
and step A, based on the step length adjustment strategy, identifying a fourth frame image containing the test question, which is before the second frame image and is separated from the second frame image by the nearest distance, and taking the position of the fourth frame image as a second temporary starting point of the test question end point searching range. The searching method of the fourth frame image may refer to the searching method of the third frame image.
In this embodiment, the position of the second frame image, which does not include the test question after the start frame image, may be represented as j; in a specific example, the second temporary start point may be denoted as s2_ s = j-step 1; it should be noted that if a position before the second frame image and spaced apart from the second frame image by a certain step exceeds the start point (denoted as 0) of the original video, the second temporary start point is determined as the start point of the original video, i.e., s2_ s = 0.
Step B, taking the position of the second frame image as a second temporary end point of the test question end point searching range; the second temporary end point may be denoted as s2_ e = j.
Step C, repeatedly performing a second adjustment operation as shown in the following 1) and 2) based on the second temporary start point and the second temporary end point:
1) determining a second temporary intermediate point between the second temporary start point and the second temporary end point, and determining a second candidate image of a frame next to the second temporary intermediate point. Specifically, the second temporary intermediate point may be represented as s2_ c = (s2_ s + s2_ e)/2, and the second candidate image may be represented as s2_ { c +1 }.
2) Respectively identifying whether the frame image and the second candidate image corresponding to the second temporary intermediate point contain test questions; referring to table 2, if the frame image corresponding to the second temporary intermediate point s2_ c and the second candidate image s2_ { c +1} both contain the test question, it indicates that the true end point of the test question should be located after the first candidate image s1_ { c +1}, in which case the position s2_ { c +1} of the second candidate image is adjusted to the new second temporary start point s2_ s; if no test question is included in the frame image corresponding to the second temporary intermediate point s2_ c and the second candidate image s2_ { c +1}, it indicates that the true end point of the test question should be located before the second temporary intermediate point s2_ c, in which case the second temporary intermediate point s2_ c is adjusted to a new second temporary end point s2_ e.
TABLE 2
| s2_c | s2_{c+1} | Adjusting the state |
| Including examination questions | Including examination questions | New s2_ s = = s2_ { c +1} |
| Do not contain examination questions | Do not contain examination questions | New s2_ e = = s2_ c |
| Including examination questions | Do not contain examination questions | Stopping the second adjustment operation |
And step D, repeatedly executing the second adjustment operation until the frame image corresponding to the second temporary intermediate point s2_ c contains the test question and the second candidate image s2_ { c +1} does not contain the test question, and stopping the second adjustment operation. The frame image corresponding to the second temporary intermediate point s2_ c contains a test question and the second candidate image s2_ { c +1} contains no test question, and this indicates that the frame image corresponding to the second temporary intermediate point s2_ c is the frame image containing the test question that is located closest to the second frame image j and is found before the second frame image j, and that the second adjustment operation can be stopped.
In the determination method of the termination frame image provided in this embodiment, the search range of the test question end point is gradually reduced by continuously adjusting the second temporary start point, the second temporary end point, and the second temporary intermediate point, so that the termination frame image containing the test question is accurately positioned, and the positioning accuracy of the test question end point is improved.
The method for acquiring a video clip provided in this embodiment may further include: after the target video segment containing the test questions is determined, generating the title of the target video segment according to a preset title generator, and adding the title to the target video segment.
In particular implementations, the title generator may generate the title of the target video segment using regular expressions, examples of which are such as: "((" example topic | exploration | topic | consolidation | training | expansion | challenge) [0-1] {0,1} [0-9] {0,1}) ".
According to the above embodiments, there is provided a method for acquiring a video clip as shown in fig. 3, including the following steps:
step S301, an original video is acquired.
Step S302, identifying whether the temporary frame image of the original video contains test questions. When the temporary frame image does not contain the test question, executing the following step S303; when the temporary frame image includes the test question, the following step S304 and subsequent steps are performed.
Step S303, determining a forward step length of the video forward based on a preset step length adjustment strategy, determining a frame image separated from the temporary frame image by the forward step length as a new temporary frame image, and returning to step S302.
Step S304, determining the temporary frame image as a first frame image, and determining a starting frame image of the video segment containing the test question based on the initial step length and the position of the first frame image.
Step S305, determining the advancing step length of the video advancing based on the step length adjusting strategy, and determining the frame image separated from the initial frame image by the advancing step length as the candidate frame image.
Step S306, identifying whether the candidate frame image contains a test question. When the candidate frame image contains the test question, executing the following step S307; when the candidate frame image does not include the test question, the following step S308 and subsequent steps are performed.
Step S307, determining a forward step of the video forward based on the first step adjustment strategy, determining a frame image separated from the candidate frame image by the forward step as a new candidate frame image, and returning to step S306.
In step S308, the candidate frame image is determined as the second frame image which does not include the test question after the start frame image.
Step S309, determining an ending frame image of the video segment containing the test question based on the initial step size or the position of the starting frame image and the position of the second frame image.
Step S310, a target video clip containing the test question is determined based on the starting frame image and the ending frame image. Meanwhile, the ending frame image is used as the starting point for re-executing the positioning operation, and the process returns to the step S302.
Step S311, judging whether the original video is finished; if not, go back to the step S302; if so, the following step S312 is performed.
Step S312, all target video segments containing the test questions in the original video are output.
In summary, according to the method for acquiring a video clip provided in this embodiment, on one hand, in the process of repeatedly performing the following positioning operation, the advance step length can be flexibly adjusted according to whether the frame image contains the test question or not through the step length adjustment strategy, and the efficiency and accuracy of acquiring the first frame image containing the test question and the second frame image not containing the test question are effectively improved. On the other hand, based on the position and the initial step length of the first frame image, the searching starting point of the test question can be continuously adjusted, the initial frame image can be more accurately positioned, the searching end point of the test question can be continuously adjusted by further utilizing the position of the initial frame image, and the positioning accuracy of the termination frame image can be simultaneously improved; therefore, the target video segment determined based on the starting frame image and the ending frame image can contain the test questions and has high-accuracy positioning. Therefore, the scheme improves the acquisition efficiency and accuracy of the target video clip.
Example two:
referring to fig. 4, a block diagram of an apparatus for acquiring a video clip is provided, the apparatus including:
an original video obtaining module 410, configured to obtain an original video;
a positioning operation repeat execution module 420, configured to repeatedly execute the following positioning operations until the positioning operation is executed to the end of the original video:
the identifying unit 422 is configured to identify a first frame image containing a test question in the original video based on a preset step size adjustment strategy; the step length adjusting strategy is a strategy for adjusting the advancing step length based on the result of whether the frame image contains the test question or not;
a start frame positioning unit 424, configured to determine a start frame image of a video segment containing a test question based on the initial step size and the position of the first frame image;
an end frame positioning unit 426, configured to identify a second frame image that does not include the test question after the start frame image based on the step size adjustment policy, and determine an end frame image of the video segment that includes the test question based on the initial step size or the position of the start frame image and the position of the second frame image; and returning to the step executed by the identification unit for a frame image subsequent to the termination frame image of the original video;
a target video clip obtaining module 430, configured to obtain all target video clips in the original video when the positioning operation is stopped; wherein the target video segment is a video segment containing a test question determined based on the start frame image and the end frame image.
The device provided by the embodiment has the same implementation principle and technical effect as the method embodiments, and for the sake of brief description, reference may be made to the corresponding contents in the method embodiments without reference to the device embodiments.
In one embodiment, there is also provided an electronic device including: a processor and a storage device; the storage means has stored thereon a computer program which, when executed by the processor, performs the method in the above-described embodiments.
Further, the present embodiment also provides a computer-readable storage medium storing a computer program for executing the above method.
The method, apparatus, and device for acquiring a video segment provided in the embodiments of the present disclosure include a computer-readable storage medium storing a program code, where instructions included in the program code may be used to execute the method described in the foregoing method embodiments, and specific implementation may refer to the method embodiments, and will not be described herein again.
It is noted that, in this document, relational terms such as "first" and "second," and the like, may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The foregoing are merely exemplary embodiments of the present disclosure, which enable those skilled in the art to understand or practice the present disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the disclosure. Thus, the present disclosure is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A method for acquiring a video clip, comprising:
acquiring an original video;
repeatedly executing the following positioning operation until the positioning operation is executed to the end of the original video, and stopping:
identifying a first frame image containing test questions in the original video based on a preset step length adjustment strategy; the step length adjustment strategy is a strategy for adjusting the advancing step length based on the result of whether the frame image contains the test question or not, and represents that: if the current frame image to be identified contains the test question, reducing the forward step length of the next frame image to be searched, and if the current frame image to be identified does not contain the test question, increasing the forward step length of the next frame image to be searched;
determining a starting frame image of a video segment containing the test question based on the initial step length and the position of the first frame image; the method comprises the following steps: determining a searching range of a starting point of the test questions based on the position of the first frame image and the initial step length, and determining a starting frame image which is most front in sequence and contains the test questions in the searching range of the starting point of the test questions; the temporary starting point of the test question starting point searching range is a third frame image which is before the first frame image and is closest to the first frame image and does not contain the test question and is identified based on the step length adjustment strategy;
identifying a second frame image which does not contain the test question after the initial frame image based on the step length adjusting strategy, and determining an ending frame image of the video segment containing the test question based on the initial step length or the position of the initial frame image and the position of the second frame image; the method comprises the following steps: determining a test question end point searching range based on the position of the second frame image and the initial step length, and determining a final frame image which is ranked most backwards and contains test questions in the test question end point searching range; the temporary starting point of the test question end point searching range is a fourth frame image which is before the second frame image and is closest to the second frame image and contains the test questions and is identified based on the step length adjusting strategy;
returning to the step of identifying the first frame image containing the test question in the original video based on a preset step length adjusting strategy aiming at the frame image behind the termination frame image in the original video;
when the positioning operation is stopped, acquiring all target video clips in the original video; wherein the target video segment is a video segment containing a test question determined based on the start frame image and the end frame image.
2. The method of claim 1, wherein the step size adjustment policy is:
wherein, a1Adjusting the step length adjustment coefficient in the strategy for the first step length, and a1＜0，a2A step size adjustment coefficient in the second step size adjustment strategy, and2＞0，stepistep for the current step sizei+1Step for the next step after the current step is adjustediAnd stepi+1Are constrained within a preset step range.
3. The method of claim 2, wherein the method comprises:
identifying whether the current frame image to be identified contains test questions;
if the frame image to be identified contains the test question, determining the next target step length based on the current step length and the first step length adjusting strategy; if the frame image to be identified does not contain the test question, determining the next target step length based on the current step length and a second step length adjusting strategy;
and determining the frame image which is separated from the frame image to be identified by the target step length as a new frame image to be identified, and returning to the step of identifying whether the current frame image to be identified contains the test question or not.
4. The method according to claim 3, wherein the identifying whether the current frame image to be identified contains the test question comprises:
carrying out feature extraction on a frame image to be identified through a feature extraction network to obtain a feature image;
recognizing text information of the characteristic image according to the optical character recognition COR, and converting the text information into word vectors;
judging whether the word vector is a test question or not according to a preset question discriminator;
and when the word vector is a test question, determining that the frame image to be identified contains the test question.
5. The method according to claim 1, wherein the identifying a first frame of image containing a test question in the original video based on a preset step size adjustment strategy comprises:
repeatedly executing the following identification operation until the temporary frame image is identified to contain the test question, and determining a first frame image containing the test question:
identifying whether a temporary frame image in the original video contains a test question or not; when the positioning operation is executed for the first time, the temporary frame image is a preset frame image of the original video; when the positioning operation is not executed for the first time, the temporary frame image is a frame image which is separated from an ending frame image determined when the positioning operation is executed for the last time by an initial step length;
if the temporary frame image does not contain the test question, determining the forward step length of the video forward based on a preset step length adjustment strategy, and determining the frame image which is separated from the temporary frame image by the forward step length as a new temporary frame image;
and if the temporary frame image contains the test question, determining the temporary frame image as a first frame image.
6. The method of claim 1, wherein determining a starting frame image of a video segment containing a test question based on the initial step size and the position of the first frame image comprises:
based on the step length adjustment strategy, identifying a third frame image which is before the first frame image and is separated from the first frame image by the nearest distance and does not contain the test question, and taking the position of the third frame image as a first temporary starting point of a test question starting point searching range; wherein the number of spaced frames between the first frame image and the third frame image is determined based on the step size adjustment policy and the initial step size;
taking the position of the first frame image as a first temporary end point of a test question starting point searching range;
based on the first temporary starting point and the first temporary ending point, repeatedly performing a first adjustment operation as follows:
determining a first temporary intermediate point between the first temporary start point and the first temporary end point, and determining a first candidate image of a frame next to the first temporary intermediate point;
respectively identifying whether the frame image corresponding to the first temporary intermediate point and the first candidate image contain test questions; if the frame image corresponding to the first temporary intermediate point and the first candidate image both contain test questions, adjusting the first temporary intermediate point to be a new first temporary end point; if the frame image corresponding to the first temporary intermediate point and the first candidate image do not contain the test question, adjusting the position of the first candidate image to be a new first temporary starting point;
and repeatedly executing the first adjustment operation until the frame image corresponding to the first temporary intermediate point does not contain the test question and the first candidate image contains the test question, stopping the first adjustment operation, and determining the first candidate image corresponding to the first adjustment operation when the first adjustment operation is stopped as the initial frame image of the video segment containing the test question.
7. The method of claim 1, wherein determining an ending frame image of the video segment containing the test question based on the initial step size and the position of the second frame image comprises:
based on the step length adjustment strategy, identifying a fourth frame image which is before the second frame image and is separated from the second frame image by the nearest test question, and taking the position of the fourth frame image as a second temporary starting point of a test question end point searching range; wherein the number of frames apart between the second frame image and the fourth frame image is determined based on the step size adjustment policy and the initial step size;
taking the position of the second frame image as a second temporary end point of the test question end point searching range;
based on the second temporary starting point and the second temporary ending point, repeatedly performing a second adjustment operation as follows:
determining a second temporary intermediate point between the second temporary start point and the second temporary end point, and determining a second candidate image of a frame next to the second temporary intermediate point;
respectively identifying whether the frame image corresponding to the second temporary intermediate point and the second candidate image contain test questions; if the frame image corresponding to the second temporary intermediate point and the second candidate image both contain the test question, adjusting the position of the second candidate image to be a new second temporary starting point; if the frame image corresponding to the second temporary intermediate point and the second candidate image do not contain the test question, adjusting the second temporary intermediate point to be a new second temporary end point;
and repeatedly executing the second adjustment operation until the frame image corresponding to the second temporary intermediate point contains the test question and the second candidate image does not contain the test question, stopping the second adjustment operation, determining the corresponding second temporary intermediate point when the second adjustment operation is stopped as a target end point, and determining the frame image corresponding to the target end point as an end frame image of the video clip containing the test question.
8. The method of claim 1, further comprising:
after a target video segment containing test questions is determined, generating a title of the target video segment according to a preset title generator, and adding the title to the target video segment.
9. An apparatus for acquiring a video clip, comprising:
the original video acquisition module is used for acquiring an original video;
a positioning operation repeated execution module, configured to repeatedly execute the following positioning operations until the positioning operations are executed until the end of the original video:
the identification unit is used for identifying a first frame image containing the test question in the original video based on a preset step length adjustment strategy; the step length adjustment strategy is a strategy for adjusting the advancing step length based on the result of whether the frame image contains the test question or not, and represents that: if the current frame image to be identified contains the test question, reducing the forward step length of the next frame image to be searched, and if the current frame image to be identified does not contain the test question, increasing the forward step length of the next frame image to be searched;
the starting frame positioning unit is used for determining a starting frame image of a video segment containing the test question based on the initial step length and the position of the first frame image; the method comprises the following steps: determining a searching range of a starting point of the test questions based on the position of the first frame image and the initial step length, and determining a starting frame image which is most front in sequence and contains the test questions in the searching range of the starting point of the test questions; the temporary starting point of the test question starting point searching range is a third frame image which is before the first frame image and is closest to the first frame image and does not contain the test question and is identified based on the step length adjustment strategy;
the termination frame positioning unit is used for identifying a second frame image which does not contain the test question after the initial frame image based on the step length adjusting strategy, and determining a termination frame image of a video segment containing the test question based on the initial step length or the position of the initial frame image and the position of the second frame image; the method comprises the following steps: determining a test question end point searching range based on the position of the second frame image and the initial step length, and determining a final frame image which is ranked most backwards and contains test questions in the test question end point searching range; the temporary starting point of the test question end point searching range is a fourth frame image which is before the second frame image and is closest to the second frame image and contains the test questions and is identified based on the step length adjusting strategy; and returning to the step executed by the identification unit for a frame image subsequent to the termination frame image of the original video;
a target video clip obtaining module, configured to obtain all target video clips in the original video when the positioning operation is stopped; wherein the target video segment is a video segment containing a test question determined based on the start frame image and the end frame image.
10. An electronic device, characterized in that the electronic device comprises: a processor and a storage device;
the storage device has stored thereon a computer program which, when executed by the processor, performs the method of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110392468.XA CN112804586B (en) | 2021-04-13 | 2021-04-13 | Method, device and equipment for acquiring video clip |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110392468.XA CN112804586B (en) | 2021-04-13 | 2021-04-13 | Method, device and equipment for acquiring video clip |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112804586A CN112804586A (en) | 2021-05-14 |
| CN112804586B true CN112804586B (en) | 2021-07-16 |
Family
ID=75816873
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110392468.XA Active CN112804586B (en) | 2021-04-13 | 2021-04-13 | Method, device and equipment for acquiring video clip |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112804586B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8644691B1 (en) * | 2012-11-29 | 2014-02-04 | Nvidia Corporation | Adaptively and accurately seeking a video frame of a video sequence rendered on a data processing device |
| CN106777159A (en) * | 2016-12-20 | 2017-05-31 | 西安交通大学 | A kind of video clip retrieval and localization method based on content |
| CN109618184A (en) * | 2018-12-29 | 2019-04-12 | 北京市商汤科技开发有限公司 | Method for processing video frequency and device, electronic equipment and storage medium |
| CN111294524A (en) * | 2020-02-24 | 2020-06-16 | 中移（杭州）信息技术有限公司 | Video editing method and device, electronic equipment and storage medium |
| CN112053399A (en) * | 2020-09-04 | 2020-12-08 | 厦门大学 | Method for positioning digestive tract organs in capsule endoscope video |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111901536B (en) * | 2020-08-04 | 2023-03-24 | 携程计算机技术（上海）有限公司 | Video editing method, system, device and storage medium based on scene recognition |
| CN112380929A (en) * | 2020-10-30 | 2021-02-19 | 北京字节跳动网络技术有限公司 | Highlight segment obtaining method and device, electronic equipment and storage medium |
-
2021
- 2021-04-13 CN CN202110392468.XA patent/CN112804586B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8644691B1 (en) * | 2012-11-29 | 2014-02-04 | Nvidia Corporation | Adaptively and accurately seeking a video frame of a video sequence rendered on a data processing device |
| CN106777159A (en) * | 2016-12-20 | 2017-05-31 | 西安交通大学 | A kind of video clip retrieval and localization method based on content |
| CN109618184A (en) * | 2018-12-29 | 2019-04-12 | 北京市商汤科技开发有限公司 | Method for processing video frequency and device, electronic equipment and storage medium |
| CN111294524A (en) * | 2020-02-24 | 2020-06-16 | 中移（杭州）信息技术有限公司 | Video editing method and device, electronic equipment and storage medium |
| CN112053399A (en) * | 2020-09-04 | 2020-12-08 | 厦门大学 | Method for positioning digestive tract organs in capsule endoscope video |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112804586A (en) | 2021-05-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8380045B2 (en) | Systems and methods for robust video signature with area augmented matching | |
| CN110769178B (en) | Method, device and equipment for automatically generating goal shooting highlights of football match and computer readable storage medium | |
| CN110321470A (en) | Document processing method, device, computer equipment and storage medium | |
| CN104766077B (en) | Method and device for recognizing characters in picture | |
| CN112382295B (en) | Speech recognition method, device, equipment and readable storage medium | |
| CN110751030A (en) | Video classification method, device and system | |
| CN110347866B (en) | Information processing method, information processing device, storage medium and electronic equipment | |
| CN107092602B (en) | Automatic response method and system | |
| CN109471955B (en) | Video clip positioning method, computing device and storage medium | |
| CN112804586B (en) | Method, device and equipment for acquiring video clip | |
| CN110795597A (en) | Video keyword determination method, video retrieval method, video keyword determination device, video retrieval device, storage medium and terminal | |
| CN111079777B (en) | Page positioning-based click-to-read method and electronic equipment | |
| CN116644228A (en) | Multi-mode full text information retrieval method, system and storage medium | |
| CN112668561B (en) | Teaching video segmentation determination method and device | |
| CN109783679B (en) | Learning auxiliary method and learning equipment | |
| CN112992148A (en) | Method and device for recognizing voice in video | |
| US9224040B2 (en) | Method for object recognition and describing structure of graphical objects | |
| CN111444386A (en) | Video information retrieval method and device, computer equipment and storage medium | |
| CN109710844A (en) | The method and apparatus for quick and precisely positioning file based on search engine | |
| CN113269141B (en) | Image processing method and device | |
| CN115422399B (en) | Video searching method, device, equipment and storage medium | |
| CN114398463B (en) | Voice tracking method and device, storage medium and electronic equipment | |
| US20050213819A1 (en) | Form recognition system, method, program, and storage medium | |
| CN113656553B (en) | Paper-electricity synchronous retrieval method, electronic equipment and computer storage medium | |
| CN111950332B (en) | Video time sequence positioning method, device, computing equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |