PPT highlight display control system based on voice recognition
Through the PPT highlighting control system based on voice recognition, the PPT content is automatically highlighted, which solves the problem that the presenter's manual operation affects the fluency of speech, and improves the demonstration efficiency and effect.
Patent Information
- Application Number
- CN202411017228.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-29
- Publication Date
- 2024-11-29
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
During the speech and presentation process, the presenter needs to manually control the highlighting of the PPT, which leads to distracting concentration, affecting the fluency and presentation quality of the speech, and also affects the viewing experience of the audience.
A PPT highlighting control system based on speech recognition is adopted, including a voice recognition module, a text analysis module and a highlighting control module. The corresponding content in the PPT is automatically highlighted through the voice content, and the voice recognition module is used to collect and convert voice , the text analysis module extracts keywords, the highlight control module generates highlight control instructions, and the PPT processing module performs highlighting processing.
实现了自动高亮显示,提升了演示效率和效果,减少了演示者手动操作的需求,提高了演讲的流畅性和观众的观看体验。
Smart Images

Figure CN119049472A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of slide presentation control, and in particular to a PPT highlight display control system based on voice recognition. Background Art
[0002] During speeches and presentations, presenters usually need to manually control the highlighted parts of the PPT so that the audience can better follow the explanation. However, manual operation requires accurately locating the content of the current progress from the numerous contents of the PPT and specifying the form of highlighting. This process will affect the presenter's concentration and thus affect the fluency of the speech. The speaker's distraction will also affect the quality of the presentation content, and will also have a certain impact on the audience's viewing experience. Summary of the invention
[0003] The purpose of the present invention is to provide a PPT highlight display control system based on voice recognition. The present invention automatically highlights the corresponding content in the PPT through the voice content, thereby improving the demonstration efficiency and effect.
[0004] The technical solution of the present invention is as follows: a PPT highlight display control system based on voice recognition, comprising a voice recognition module, a text analysis module, a highlight display control module and a PPT processing module; the voice recognition module is used to collect voice and convert it into text; the text analysis module is used to segment words and extract keywords to be highlighted in the text; the highlight display control module generates highlight display control instructions for corresponding content based on the keywords; the PPT processing module is used to receive the highlight display control instructions, and parse the PPT and perform highlight display processing on the corresponding content based on the instructions.
[0005] In the above-mentioned PPT highlight display control system based on voice recognition, the process of the voice recognition module collecting voice and converting it into text is carried out according to the following steps:
[0006] Step S1: audio collection;
[0007] Step S2: performing noise filtering, echo cancellation and gain control on the collected audio;
[0008] Step S3: Select a sampling rate of 16kHz or 44.1kHz and use a 16-bit quantization bit number to sample the audio of step S2, balancing audio quality and processing efficiency;
[0009] Step S4: The audio of step S3 is recognized by a speech recognition engine. During the recognition process, the audio is processed through feature extraction, acoustic model and language model to obtain recognized text.
[0010] In the aforementioned PPT highlighting control system based on speech recognition, the process of extracting keywords to be highlighted in the text by the text analysis module is performed according to the following steps:
[0011] Step A1: Perform word segmentation on the text through natural language processing to divide the continuous text into independent words;
[0012] Step A2: construct a stop word list, and filter the stop words in the segmented text according to the stop word list;
[0013] Step A3: Extract keywords from the text from multiple dimensions through TF-IDF, LDA topic model and named entity recognition model.
[0014] In the aforementioned PPT highlight display control system based on speech recognition, the TF-IDF is calculated by calculating the word frequency TF, inverse text word frequency IDF and TF-IDF value of each word, and taking multiple words with high TF-IDF values as keywords. The formula is as follows:
[0015]
[0016] TF-IDF(t)=TF(t)×IDF(t).
[0017] In the aforementioned PPT highlight display control system based on voice recognition, the process of the highlight display control module generating the highlight display control instruction of the corresponding content is performed according to the following steps:
[0018] Step B1: Search for keywords in the PPT document and locate the specific position of the keywords in the PPT through a string matching algorithm;
[0019] Step B2: Extract the context information before and after the keyword, and analyze the semantic relationship of the context through NLP technology; if the context is semantically coherent, proceed to step B4; if the context exists in information irrelevant to the keyword and interference information, proceed to step B3;
[0020] Step B3: After re-adjusting the context extraction range, perform semantic analysis. If there is still irrelevant information and interference information, feedback the analysis error to the user. After the user manually intervenes and corrects it, proceed to step B4;
[0021] Step B4: Convert the keyword into a highlight display control instruction with the keyword, slide number, keyword starting position, keyword length, highlight color and context content. In the aforementioned PPT highlight display control system based on speech recognition, the process of highlight display processing by the PPT processing module is carried out according to the following steps:
[0022] Step C1: parsing the highlight display control instruction to extract the information of keywords, slide number, starting position, length and color;
[0023] Step C2: locating the target text in the designated slide according to the key words, starting position, slide number and length in the highlight control instruction;
[0024] Step C3: setting the text color and background color of the target text according to the color information in the highlight display control instruction;
[0025] Step C4: Update the slide content.
[0026] In the aforementioned PPT highlight display control system based on voice recognition, testing and verification are performed after the slide content is updated, and status information of success or failure of the execution is fed back to the highlight display control module based on the results.
[0027] In the aforementioned PPT highlight display control system based on speech recognition, the verification and testing process is carried out according to the following steps:
[0028] Step C5.1: Perform a visual check on the PPT document to confirm that the keywords are highlighted correctly and that the colors and positions are correct;
[0029] Step C5.2: Test whether the highlighting effect is stable during the PPT presentation;
[0030] Step C5.3: Use the example text to confirm the semantic coherence and accuracy of the highlighted content;
[0031] Step C5.4: Simulate abnormal conditions of network interruption, command loss, or erroneous command transmission.
[0032] In the aforementioned PPT highlight control system based on voice recognition, the highlight control module encapsulates the highlight control instructions into the network transmission protocol and transmits them to the PPT processing module through encryption technology, and uses a retransmission mechanism and error detection technology to ensure the reliability of instruction transmission.
[0033] Compared with the prior art, the present invention has the following beneficial effects:
[0034] The speech recognition module of the present invention obtains the speech content of the presenter and converts it into text content, the text analysis module performs word segmentation processing on the text content and extracts the keywords to be highlighted, the highlight control module generates a highlight control instruction for the corresponding content according to the keywords of the text analysis module, and the PPT processing module receives the highlight control instruction and performs highlight processing on the corresponding content in the PPT; the present invention can automatically highlight the corresponding content in the PPT according to the speech content, thereby improving the demonstration efficiency and effect. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] Figure 1 It is a schematic diagram of the system architecture of the present invention. DETAILED DESCRIPTION
[0036] The present invention is further described below in conjunction with the accompanying drawings and embodiments, but they are not intended to limit the present invention.
[0037] Embodiment: A PPT highlight control system based on speech recognition, comprising a speech recognition module, a text analysis module, a highlight control module and a PPT processing module; the speech recognition module is used to collect speech and convert it into text; the text analysis module is used to segment words and extract keywords to be highlighted in the text; the highlight control module generates highlight control instructions for corresponding content based on the keywords; the PPT processing module is used to receive the highlight control instructions, parse the PPT according to the instructions, and perform highlight processing on the output device for the corresponding content.
[0038] The process of the speech recognition module collecting speech and converting it into text is carried out in the following steps:
[0039] Step S1: Use a high-sensitivity, low-noise microphone to collect audio to ensure clear audio input;
[0040] Step S2: noise filtering, echo elimination and gain control are performed on the collected audio; noise filtering uses bandpass filtering, Kalman filtering or spectral subtraction techniques to reduce the interference of environmental noise on the voice signal; echo elimination uses an echo elimination algorithm to detect and eliminate echoes to improve the clarity of the voice signal; gain control adjusts the strength of the audio signal through automatic gain control (AGC) to ensure that the audio signal strength is moderate;
[0041] Step S3: Select a sampling rate of 16kHz or 44.1kHz and use a 16-bit quantization bit number to sample the audio of step S2, balancing audio quality and processing efficiency;
[0042] Step S4: The audio signal of step S3 is recognized by a speech recognition engine (such as Google Speech-to-Text, Microsoft AzureSpeech). During the recognition process, the audio signal is processed by feature extraction (such as MFCC, PLP), acoustic model (such as HMM-DNN, end-to-end model) and language model (such as N-gram, RNN, Transformer) to obtain a highly accurate recognized text.
[0043] The process of extracting keywords to be highlighted in the text by the text analysis module is performed as follows:
[0044] Step A1: Segment the text through natural language processing (including forward maximum matching (FMM), reverse maximum matching (RMM), bidirectional maximum matching (BMM), and statistical-based word segmentation methods), and split the continuous text into independent words;
[0045] Step A2: Construct a stop word list (including words with little semantic contribution such as "de", "shi", "zai", etc.), and filter the stop words from the segmented text according to the stop word list;
[0046] Step A3: Extract keywords from the text from multiple dimensions through TF-IDF, LDA topic model, and named entity recognition model;
[0047] The TF-IDF calculates the term frequency TF, inverse document frequency IDF (the inverse frequency of the word in the entire text), and TF-IDF value of each word, and uses multiple words with high TF-IDF values as keywords. The formula is as follows:
[0048]
[0049] TF-IDF(t) = TF(t) × IDF(t);
[0050] The LDA topic model performs topic modeling on the text and extracts topic words related to the speech content;
[0051] The named entity recognition model uses pre-trained models (such as BERT, BiLSTM-CRF) to identify specific entities such as person names, place names, and organization names in the text;
[0052] The process of the highlighting control module generating the highlighting control instruction for the corresponding content is carried out according to the following steps:
[0053] Step B1: Search for keywords in the PPT document, and locate the specific positions of the keywords in the PPT through string matching algorithms (such as KMP, BM algorithms);
[0054] Step B2: Extract the context information before and after the keywords, including a certain number of words or sentences before and after the keywords, and analyze the semantic relationship of the context through NLP techniques, including part-of-speech tagging, entity recognition, and syntactic analysis, to understand the semantic relationship and logical structure between words in the text; if the context is semantically coherent, go to Step B4; if there is information and interference information unrelated to the keywords in the context, such as misidentified relevant words or phrases, go to Step B3;
[0055] Step B3: After readjusting the context extraction scope, perform semantic analysis, or introduce more complex semantic models and algorithms to improve accuracy. If there is still irrelevant information and interference information, feedback the analysis error to the user, provide correction suggestions or operation options, and proceed to step B4 after the user manually intervenes and corrects it;
[0056] For example, if the keyword is "AI", the context extraction range is the sentence before and after the keyword:
[0057] Text paragraph: "Artificial Intelligence (AI) is transforming industriesworldwide."
[0058] The context extracted by the system includes "Artificial Intelligence (AI) is transforming industries". Through NLP technology analysis, the system understands that there is a correlation between the semantic relationship between "AI" and "transforming industries".
[0059] Step B4: Use the PPTXGenJS library to convert keywords into highlighting control instructions in JSON format that contain keywords, slide numbers, keyword starting positions, keyword length, highlight colors, and context content; for example, to highlight the keyword "AI", the instruction is:
[0060]
[0061] The highlight display control module encapsulates the JSON format highlight display control instructions into the network transmission protocol (HTTP POST) through the network interface (HTTP) and transmits it to the PPT processing module through encryption technology (such as HTTPS, SSL / TLS), using a retransmission mechanism and error detection technology to ensure the reliability of instruction transmission.
[0062] The process of highlighting processing by the PPT processing module is carried out according to the following steps:
[0063] Step C1: Parse the highlight display control command and extract the keyword, slide number, start position, length and color information according to the JSON format;
[0064] Step C2: Use the python-pptx library to parse the PPT file, and locate the target text in the specified slide through a string matching algorithm (such as the KMP algorithm and the BM algorithm) according to the keywords, starting position, slide number and length in the highlight control instruction;
[0065] Step C3: setting the text color and background color of the target text according to the color information in the highlight display control instruction;
[0066] Step C4: Update the slide content to ensure that the highlighting effect takes effect immediately during the presentation;
[0067] Step C5: Perform testing and verification to ensure that the highlight display effect meets expectations, and feedback status information of success or failure to the highlight display control module based on the results.
[0068] The verification and testing process is carried out in the following steps:
[0069] Step C5.1: Perform a visual check on the PPT document using a visual engine to confirm whether the keywords are correctly highlighted and whether the colors and positions are correct;
[0070] Step C5.2: Test whether the highlighting effect is stable during the transition between slides in the PPT presentation;
[0071] Step C5.3: Use example text or actual demonstration content to confirm the semantic coherence and accuracy of the highlighted content;
[0072] Step C5.4: Simulate abnormal situations such as network interruption, command loss or erroneous command transmission to verify the system's fault tolerance and recovery mechanism;
[0073] If verification and testing confirm that the highlighting effect meets expectations, status information of successful execution is sent to the highlighting control module, including execution time and operation log, for further analysis and optimization. If verification and testing find any problems or situations that do not meet expectations, status information of failed execution is sent to the highlighting control module, including reasons for failure, error codes, and specific failed operations. Corresponding feedback and suggestions are provided to users or administrators to help solve problems or adjust operations.
[0074] In summary, the present invention combines speech recognition with PPT highlighting display. The speech recognition presenter's sound signal is converted into high-accuracy text. The text analysis module combines TF-IDF, LDA topic model and NER technology to extract keywords from multiple dimensions, thereby improving the accuracy and comprehensiveness of keyword extraction. The highlighting control module further performs context analysis in the PPT document based on the keywords and converts them into executable highlighting control instructions to ensure the semantic accuracy and coherence of the highlighted content. The PPT processing module uses the PPT parsing library to read the PPT file content, and dynamically modifies the style of the text in the PPT according to the highlighting control instructions to achieve a highlighting effect. The present invention automatically highlights the corresponding content in the PPT through the speech content, thereby improving the presentation efficiency and effect.
Claims
1. A PPT highlight display control system based on speech recognition, characterized in that: It includes a speech recognition module, a text analysis module, a highlight display control module and a PPT processing module; the speech recognition module is used to collect speech and convert it into text; the text analysis module is used to segment words and extract keywords to be highlighted in the text; the highlight display control module generates highlight display control instructions for corresponding content based on keywords; the PPT processing module is used to receive highlight display control instructions, and parse the PPT according to the instructions and perform highlight display processing on the corresponding content.
2. The PPT highlight display control system based on voice recognition according to claim 1, characterized in that: The process of the speech recognition module collecting speech and converting it into text is carried out in the following steps: Step S1: audio collection; Step S2: performing noise filtering, echo cancellation and gain control on the collected audio; Step S3: Select a sampling rate of 16kHz or 44.1kHz and use a 16-bit quantization bit number to sample the audio of step S2, balancing audio quality and processing efficiency; Step S4: The audio of step S3 is recognized by a speech recognition engine. During the recognition process, the audio is processed through feature extraction, acoustic model and language model to obtain recognized text.
3. The PPT highlight display control system based on voice recognition according to claim 1, characterized in that: The process of extracting keywords to be highlighted in the text by the text analysis module is performed as follows: Step A1: Perform word segmentation on the text through natural language processing to divide the continuous text into independent words; Step A2: construct a stop word list, and filter the stop words in the segmented text according to the stop word list; Step A3: Extract keywords from the text from multiple dimensions through TF-IDF, LDA topic model and named entity recognition model.
4. The PPT highlight display control system based on voice recognition according to claim 1, characterized in that: The TF-IDF is calculated by calculating the word frequency TF, inverse text word frequency IDF and TF-IDF value of each word, and taking multiple words with high TF-IDF values as keywords. The formula is as follows: TF-IDF(t)=TF(t)×IDF(t).
5. The PPT highlight display control system based on voice recognition according to claim 1, characterized in that: The process of generating the highlight display control instruction corresponding to the content by the highlight display control module is performed according to the following steps: Step B1: Search for keywords in the PPT document and locate the specific position of the keywords in the PPT through a string matching algorithm; Step B2: Extract the contextual information before and after the keywords and analyze the semantic relationship of the context using NLP technology; If the context is semantically coherent, proceed to step B4; If the context exists in keyword-irrelevant information and interference information, proceed to step B3; Step B3: After re-adjusting the context extraction range, perform semantic analysis. If there is still irrelevant information and interference information, feedback the analysis error to the user. After the user manually intervenes and corrects it, proceed to step B4; Step B4: Convert the keywords into highlighting control instructions containing the keywords, slide numbers, keyword starting positions, keyword lengths, highlight colors and contextual contents.
6. The PPT highlight display control system based on voice recognition according to claim 5, characterized in that: The process of highlighting processing by the PPT processing module is carried out according to the following steps: Step C1: parsing the highlight display control instruction to extract the keyword, slide number, start position, length and color information; Step C2: locating the target text in the designated slide according to the key words, starting position, slide number and length in the highlight control instruction; Step C3: setting the text color and background color of the target text according to the color information in the highlight display control instruction; Step C4: Update the slide content.
7. The PPT highlight display control system based on voice recognition according to claim 6, characterized in that: After updating the slide content, test and verify, and feedback the status information of success or failure of execution to the highlight control module based on the results.
8. The PPT highlight display control system based on voice recognition according to claim 7, characterized in that: The verification and testing process is carried out in the following steps: Step C5.1: Perform a visual check on the PPT document to confirm that the keywords are highlighted correctly and that the colors and positions are correct; Step C5.2: Test whether the highlighting effect is stable during the PPT presentation; Step C5.3: Use the example text to confirm the semantic coherence and accuracy of the highlighted content; Step C5.4: Simulate abnormal conditions of network interruption, command loss, or erroneous command transmission.
9. The PPT highlight display control system based on voice recognition according to claim 1, characterized in that: The highlight display control module encapsulates the highlight display control instruction into the network transmission protocol and transmits it to the PPT processing module through encryption technology, and uses a retransmission mechanism and error detection technology to ensure the reliability of instruction transmission.
Citation Information
Patent Citations
Demonstration document operation method, demonstration document operation system, demonstration document control device and demonstration document demonstration device
CN105635831A
Display method and apparatus
CN105808733A
Demonstration document special effect matching method and apparatus
CN107544985A
Information processing apparatus, information processing system, and information processing method
CN107766321A
Generation method and device of presentation document and equipment
CN108509405A
Cited By
A method for dynamically managing and navigating associated content in a presentation system, apparatus, and computer-program product
WO2026137309A1
