基于语音识别技术的动画资源生成研究

打开文本图片集

中图分类号：TP391 文献标识码：A 文章编号：1672-3791（2026）03-0025-05

Research on GenerationofAnimation ResourcesBased on Speech Recognition Technology

KANG Dong

Hainan College of Economicsand Business，Haikou，Hainan Province，571127 China

Abstract： Based on the structure of speech recognition technology，ananimation generation technology path based on speech recognition is proposed by taking mouth animation as an example.After the feature extraction of the speech，an acoustic modelis used to generate phoneme sequences，and the pronunciation of each word in the phoneme sequence is mapped and retrieved to generate initial anmations from existing animation resource libraries.Finaly，the video is optimizedand edited according to the slice duration of the original sound，so as to generate an animation result.Experimental results show that compared to lip animations created using traditional methods，lip animations generated based onspeechrecognition technology exhibit significantadvantages inaccuracy，naturalne and fluency，and anomaly statistics.

eywords： Animation generation; Lip-sync animation; Speech recognition; Phoneme sequence

人工智能驱动数字经济飞速发展的今天，数字化动画资源市场呈现前所未有的繁荣景象，其需求量呈爆发式增长态势。（剩余5165字）