Method for archiving and restoring the relationship between transcripts and video recordings and reading DLF files

Abstract

The present invention discloses a method for archiving and restoring relational reading of a DLF file generated from a transcript file and a video recording, comprising the following steps: obtaining original data of the transcript file and the video recording; generating a mapping relationship table between the transcript file and the video time axis; performing structured processing on the transcript file to generate a transcript metadata file in XML format; creating a DLF file container; dividing the video file into a key frame sequence to generate a dynamic video OFD component; embedding the structured transcript file into an extension layer of the dynamic video OFD component to establish a spatiotemporal correlation index between the transcript text and the video key frames; registering dynamic interactive events in a navigation file to construct a multi-layer nested structure; generating an interactive directory tree, automatically creating a navigable bookmark identifier based on transcript paragraph keywords, and enabling dynamic retrieval; and performing digital signature verification and archiving on the combined DLF file, thereby realizing relational reading in an archive system.

Description

Technical Field

[0001] The invention belongs to the technical field of files, and in particular relates to a method for archiving and restoring a DLF file generated by a transcript file and a video. Background Art

[0002] During the trial, the trial process needs to be videotaped and recorded. At the same time, the record files and video files need to be archived in a timely manner. During the archiving process, they are archived in the form of layout files according to the national standard GB / T 33190-2016 "Electronic Document Storage and Exchange Format Layout Document". However, after archiving, the files lose their original application system environment and cannot be associated with reading, resulting in a low reading experience for users. Summary of the Invention

[0003] The purpose of the present invention is to provide a method for archiving and restoring the relationship between DLF files generated by transcripts and videos, so as to solve the problems raised in the above background technology.

[0004] To achieve the above object, the present invention provides the following technical solution: a method for archiving and restoring the relationship between recorded files and video files and reading the DLF file, comprising the following steps:

[0005] S1: Obtaining original data of a transcript file and a video recording. The transcript file format includes at least one of Word, TXT, SRT, and XML, and the video recording format includes at least one of MP4, AVI, and MOV.

[0006] S2: Use speech recognition technology to convert the audio stream in the video into time-stamped text data and generate a mapping table between the transcript file and the video timeline;

[0007] S3: Structural processing of the transcript file, extraction of timestamps, speaker identifiers, and semantic paragraphs, and generation of transcript metadata files in XML format;

[0008] S4: Create a DLF file container, and create a navigation file, a video resource directory, a transcript resource directory, and associated index files in the compressed package;

[0009] S5: Segment the video file into key frame sequences, encapsulate the video key frames and timeline data using the OFD standard, and generate dynamic video OFD components;

[0010] S6: embed the structured transcript file into the extension layer of the dynamic video OFD component and establish a spatiotemporal correlation index between the transcript text and the video key frames;

[0011] S7: Register dynamic interaction events in the navigation file to automatically jump to the corresponding video time node when clicking on the transcript segment;

[0012] S8: Construct a multi-layer nested structure, with the transcript text layer, video frame layer, and timeline layer using independent and switchable display levels;

[0013] S9: Generate an interactive directory tree, automatically create navigable bookmarks based on the keywords of the transcript paragraphs, and enable dynamic retrieval;

[0014] S10: Perform digital signature verification on the combined DLF file, save the file hash value and version metadata. The digital signature verification uses a national secret algorithm to generate a 256-bit hash value. The verification process includes file integrity verification, timestamp verification, and issuer certificate chain verification, and saves the metadata.

[0015] S11: Archiving the combined DLF file, and retaining the association relationship between the transcript file and the video file and the dynamic operation function, to achieve associated reading in the archive system.

[0016] Preferably, the time axis mapping relationship table in step S2 uses double-precision floating point numbers to record millisecond timestamps, and the synchronization error between the transcript text and the video frame does not exceed ±50ms. When the timing deviation between the transcript text and the video frame exceeds ±50ms, automatic recalibration is triggered.

[0017] Preferably, the transcript file in step S3 includes text input, voice input, file import, and image import; wherein the image import extracts the text content through OCR recognition and converts it into a structured transcript file content file, automatically segments the text according to semantics and context, and generates corresponding paragraphs in a logical order, while generating timestamps for each paragraph and marking key events. The text content is reorganized using natural language processing technology, automatically correcting spelling errors and providing completion.

[0018] Preferably, the generation of the dynamic video OFD component in step S5 includes the following steps: step a: extracting video I frames as a key frame sequence, and saving each frame as an independent OFD page; step b: creating a timeline description file, recording the timestamp and frame hash value corresponding to the key frame; step c: using H.265 encoding to perform secondary compression on the video stream, with a compression ratio of not less than 50:1.

[0019] Preferably, the spatiotemporal correlation index of step S6 is implemented through four-dimensional coordinates, including four-tuple positioning data of X-axis coordinates, Y-axis coordinates, timestamp, and transcript paragraph ID.

[0020] Preferably, the navigation file adopts an improved XML structure, including a video resource path mapping table, a transcript paragraph jump event trigger, and multi-level display control parameters.

[0021] Preferably, the generation of the interactive directory tree uses NLP keyword extraction technology to extract entity keywords in the transcript paragraphs, and automatically constructs a tree directory structure based on the OFD file association relationship in the DLF navigation file, supporting nesting of more than three levels, and each node is bound to the corresponding OFD file path and starting page number.

[0022] Preferably, the dynamic interactive event of the navigation file aligns the video key frame timestamp with the transcript text paragraph, establishes a bidirectional jump index, and accurately locates the approval area through the OFD layout coordinates (X, Y, Width, Height), so as to synchronize the highlighting of the video progress bar and the corresponding approval layer when clicking the bookmark.

[0023] Preferably, the multi-layer nested structure supports dynamic loading, including a basic display layer: presenting static images of video key frames, an enhanced interaction layer: superimposing transcript text and semantic annotations, and a timeline control layer: providing a playback control bar with an accuracy of 0.1 seconds.

[0024] Preferably, the dynamic search includes keywords, events, images, voices, and time ranges; and keyword search is performed by analyzing the semantics of the content of the transcript file through NLP; the search results are presented in a timeline and thumbnails; and the search results can be sorted according to relevance and chronological order.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] The present invention solves the problem that after a case is archived, the transcripts of law enforcement agencies can be archived in layout files according to the national standard GB / T 33190-2016 "Electronic File Storage and Exchange Format Layout Document", but the original application system environment is lost after archiving, and the associated reading cannot be carried out. Based on the DLF dynamic layout file technology, the relationship between the archived OFD file and the video file and the dynamic operation function are retained, so that readers can restore the relationship between the transcript and the video, and realize associated reading in the archive system.

[0027] The present invention generates a dynamic layout file (DLF) by combining the archived OFD file and the video, thereby ensuring the authenticity and reliability of the archived data. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 is a flow chart of the present invention; DETAILED DESCRIPTION

[0029] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention. Example

[0030] During a court trial, the video and transcript files of the trial are usually recorded separately. Later, the transcript files need to be integrated with the video files and the corresponding files need to be archived. At the same time, in order to facilitate subsequent related reading, the present invention provides a method for archiving and restoring the relationship between transcript files and video recordings by generating DLF files, including the following steps: S1: obtaining the original data of the transcript files and video recordings, the transcript file formats include Word, TXT, SRT, XML, etc., and the video recording formats include MP4, AVI, MOV, etc.;

[0031] S2: Process the transcripts and video recordings according to their different formats. Then, use speech recognition technology to convert the audio stream in the video into time-stamped text data, and generate a mapping table between the transcripts and the video timeline. The timeline mapping table uses double-precision floating-point numbers to record millisecond-level timestamps. The transcripts can come from multiple conversation files, audio recordings, etc. At the same time, the synchronization error between the transcript text and the video frame does not exceed ±50ms. If the timing deviation between the transcript text and the video frame exceeds ±50ms, automatic recalibration is triggered to ensure an accurate match between the transcript file and the video.

[0032] S3: After the audio in the video is converted into transcript text, the transcript file needs to be processed and structured. The formats of the transcript file include Word, TXT, SRT, XML and other formats. The transcript file is processed according to different formats, and the timestamp mark, speaker identifier, and semantic paragraph are extracted to generate the transcript metadata file in XML format. The input methods of the transcript file include text input, voice input, file import, and image import. The image import is to extract the text content through OCR recognition and convert it into a structured transcript file content file, and automatically segment the text according to semantics and context, and generate corresponding paragraphs in a logical order. At the same time, a timestamp is generated for each paragraph, and key events are marked. The text content is reorganized through natural language processing technology, and spelling errors are automatically corrected and completion is provided.

[0033] S4: Create a DLF file container, and create a navigation file, a video resource directory, a transcript resource directory, and associated index files in the compressed package; the navigation file refers to the directory outline, the video resource directory is the directory of various video source files; the transcript resource directory is the directory of various file contents;

[0034] S5: After creating the DLF file container, the video file is divided into a key frame sequence, and the video key frames and timeline data are encapsulated using the OFD standard to generate a dynamic video OFD component. The generation of the dynamic video OFD component includes the following steps: Step a: Extract the video I frame as a key frame sequence, and then save each frame as an independent OFD page; Step b: Create a timeline description file to record the timestamp and frame hash value corresponding to the key frame; Step c: Use H.265 encoding to perform secondary compression on the video stream, with a compression ratio of not less than 50:1. Through the above steps, the video file can be generated into an OFD component.

[0035] S6: After the OFD component is generated from the dynamic video file, the structured transcript file is embedded in the extension layer of the dynamic video OFD component to establish a spatiotemporal correlation index between the transcript text and the video key frames. The spatiotemporal correlation index is implemented through four-dimensional coordinates, which contains four-tuple positioning data of X-axis coordinate, Y-axis coordinate, timestamp, and transcript segment ID;

[0036] S7: Register dynamic interaction events in the navigation file. The dynamic interaction events in the navigation file align the video keyframe timestamps with the transcript text paragraphs, establish a bidirectional jump index, and accurately locate the approval area through the OFD layout coordinates (X, Y, Width, Height). When clicking a bookmark, the video progress bar and the corresponding approval layer are highlighted synchronously. In actual operation, clicking a transcript paragraph can automatically jump to the corresponding video time node;

[0037] S8: Build a multi-layer nested structure, where independently switchable display layers include the transcript text layer, video frame layer, and timeline layer. This multi-layer nested structure supports dynamic loading, for example, the basic display layer presents static images of video keyframes; the enhanced interaction layer overlays transcript text and semantic annotations; and the timeline control layer provides a playback control bar with 0.1-second accuracy.

[0038] S9: Generates an interactive directory tree, automatically creates navigable bookmarks based on transcript keywords, and enables dynamic retrieval; dynamic retrieval includes keywords, events, images, voice, and time ranges; and uses NLP to analyze the semantics of transcript content for keyword retrieval. Retrieval results are presented in timelines and thumbnails; and after retrieval, results can be sorted by relevance and chronological order.

[0039] S10: Perform digital signature verification on the combined DLF file, save the file hash value and version metadata. The digital signature verification uses the national secret algorithm to generate a 256-bit hash value. The verification process includes file integrity verification, timestamp verification, and issuer certificate chain verification, and saves the metadata.

[0040] S11: Archiving the combined DLF file, and retaining the association relationship between the transcript file and the video file and the dynamic operation function, to achieve associated reading in the archive system.

[0041] This invention solves the problem of law enforcement agencies losing their original application system environment and being unable to read linked documents after case filing, as they can be archived in layout files according to the national standard GB / T 33190-2016, "Electronic Document Storage and Exchange Format Layout Document." By using DLF dynamic layout file technology, the relationship between archived OFD files and video files, as well as dynamic operation functions, is preserved, allowing readers to restore the relationship between the transcript and video, thus enabling linked reading within the archive system.

[0042] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments are to be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be encompassed therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A method for archiving and restoring the relationship between transcripts and video recordings to generate DLF files, characterized in that: The following steps are involved: S1: Obtaining original data of a transcript file and a video recording. The transcript file format includes at least one of Word, TXT, SRT, and XML, and the video recording format includes at least one of MP4, AVI, and MOV. S2: Use speech recognition technology to convert the audio stream in the video into time-stamped text data and generate a mapping table between the transcript file and the video timeline; S3: Structural processing of the transcript file, extraction of timestamps, speaker identifiers, and semantic paragraphs, and generation of transcript metadata files in XML format; S4: Create a DLF file container, and create a navigation file, a video resource directory, a transcript resource directory, and associated index files in the compressed package; S5: Segment the video file into key frame sequences, encapsulate the video key frames and timeline data using the OFD standard, and generate dynamic video OFD components; S6: embed the structured transcript file into the extension layer of the dynamic video OFD component and establish a spatiotemporal correlation index between the transcript text and the video key frames; S7: Register dynamic interaction events in the navigation file to automatically jump to the corresponding video time node when clicking on the transcript segment; S8: Construct a multi-layer nested structure, with the transcript text layer, video frame layer, and timeline layer using independent and switchable display levels; S9: Generate an interactive directory tree, automatically create navigable bookmarks based on the keywords of the transcript paragraphs, and enable dynamic retrieval; S10: Perform digital signature verification on the combined DLF file, save the file hash value and version metadata. The digital signature verification uses a national secret algorithm to generate a 256-bit hash value. The verification process includes file integrity verification, timestamp verification, and issuer certificate chain verification, and saves the metadata. S11: Archiving the combined DLF file, and retaining the association relationship between the transcript file and the video file and the dynamic operation function, to achieve associated reading in the archive system.

2. The method for archiving and restoring a DLF file generated by a transcript and a video according to claim 1 is characterized in that: In step S2, the time axis mapping relationship table uses double-precision floating point numbers to record millisecond timestamps, and the synchronization error between the transcript text and the video frame does not exceed ±50ms. When the timing deviation between the transcript text and the video frame exceeds ±50ms, automatic recalibration is triggered.

3. The method for archiving and restoring DLF files generated by transcripts and videos according to claim 1 is characterized in that: The input method of the transcript file in step S3 includes text input, voice input, file import, and picture import; The image import uses OCR recognition to extract the text content and convert it into a structured transcript file. The text is automatically segmented according to semantics and context, and corresponding paragraphs are generated in a logical order. At the same time, a timestamp is generated for each paragraph and key events are marked. The text content is reorganized through natural language processing technology, and spelling errors are automatically corrected and completion is provided.

4. The method for archiving and restoring a DLF file generated by a transcript and a video according to claim 1 is characterized in that: The generation of the dynamic video OFD component in step S5 includes the following steps: step a: extracting video I frames as a key frame sequence, and saving each frame as an independent OFD page; step b: creating a timeline description file, recording the timestamps and frame hash values ​​corresponding to the key frames; Step c: Use H.265 encoding to perform secondary compression on the video stream, with a compression ratio of no less than 50:

1.

5. The method for archiving and restoring DLF files generated by transcripts and videos according to claim 1 is characterized in that: The spatiotemporal correlation index of step S6 is implemented through four-dimensional coordinates, including four-tuple positioning data of X-axis coordinates, Y-axis coordinates, timestamp, and transcript paragraph ID.

6. The method for archiving and restoring DLF files generated by transcripts and videos according to claim 1 is characterized in that: The navigation file adopts an improved XML structure, including a video resource path mapping table, a transcript paragraph jump event trigger, and multi-level display control parameters.

7. The method for archiving and restoring DLF files generated by transcripts and videos according to claim 1 is characterized in that: The interactive directory tree is generated by using NLP keyword extraction technology to extract entity keywords in the transcript paragraphs, and automatically constructs a tree directory structure based on the OFD file association relationship in the DLF navigation file, supporting nesting of more than three levels, and each node is bound to the corresponding OFD file path and starting page number.

8. The method for archiving and restoring DLF files generated by transcripts and videos according to claim 1 is characterized in that: The dynamic interactive event of the navigation file aligns the video key frame timestamp with the transcript text paragraph, establishes a bidirectional jump index, and accurately locates the signature area through the OFD layout coordinates (X, Y, Width, Height), so as to synchronize the highlighting of the video progress bar and the corresponding signature layer when clicking the bookmark.

9. The method for archiving and restoring DLF files generated by transcripts and videos according to claim 1 is characterized in that: The multi-layer nested structure supports dynamic loading, including a basic display layer: presenting static images of video key frames, an enhanced interaction layer: superimposing transcript text and semantic annotations, and a timeline control layer: providing a playback control bar with an accuracy of 0.1 seconds.

10. The method for archiving and restoring DLF files generated by transcripts and videos according to claim 1 is characterized in that: The dynamic search includes keywords, events, images, voices, and time ranges; and keyword search is performed by analyzing the semantics of the transcript file content through NLP; the search results are presented in a timeline and thumbnails; and the search results can be sorted according to relevance and chronological order.

Images