Method, system and device for automatically scrolling subtitles based on voice rhythm
A technology of scrolling subtitles and voice, applied in the field of automatic scrolling subtitles, can solve problems such as increasing costs
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0069] A method for automatically scrolling subtitles based on voice rhythm, such as figure 1 shown, including the following steps:
[0070] S100. Acquire the voice data of the subtitles to be rolled, and perform sentence segmentation processing when the voice blank gap time exceeds a preset value;
[0071] S200. Perform frame-by-frame extraction feature processing on the voice data after sentence segmentation processing, so that each frame has a fixed length, and convert each frame into a feature vector;
[0072] S300. Input the feature vector into the speech recognition model, and convert the speech data after sentence segmentation processing into human voice speech feature symbols;
[0073] S400. Comparing the feature symbols of the human voice with the preset feature symbols extracted from the subtitle text, if the similarity is higher than the threshold, the matching is successful;
[0074] S500. Dynamically calculate a smooth scrolling speed, and scroll the correspondi...
Embodiment 2
[0122] A system for automatic subtitle scrolling based on voice rhythm, such as figure 2 As shown, it includes a speech acquisition module 100, a speech extraction module 200, a speech feature conversion module 300, a feature symbol matching module 400 and a text scrolling module 500;
[0123] The voice acquisition module 100 is used to acquire the voice data of the subtitles to be scrolled, and when the voice blank gap time exceeds a preset value, sentence segmentation processing is performed;
[0124] The speech extraction module 200 is used to perform frame extraction feature processing on the speech data after sentence segmentation processing, so that each frame is a fixed length, and each frame is converted into a feature vector;
[0125] The speech feature conversion module 300 is used to input the feature vector into the speech recognition model, and convert the speech data after sentence segmentation processing into human voice speech feature symbols;
[0126] The fe...
Embodiment 3
[0140] A computer-readable storage medium, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following method steps are implemented:
[0141] Obtain the voice data of the subtitles to be rolled, and perform sentence segmentation processing when the voice blank gap time exceeds the preset value;
[0142] The speech data after sentence segmentation processing is subjected to frame-by-frame extraction feature processing, so that each frame is a fixed length, and each frame is converted into a feature vector;
[0143] Input the feature vector into the speech recognition model, and convert the speech data after sentence segmentation processing into human voice speech feature symbols;
[0144] Comparing the feature symbols of the vocal speech with the preset feature symbols extracted from the subtitle text, if the similarity is higher than the threshold, the matching is successful;
[0145] Dynamically calc...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


