Abstract | ||
---|---|---|
This paper designs a prediction model based on a Chinese talking avatar and aims to analyze the relationship between the text-based prosodic features and the nodding of head movements. To get a better predicted effect, we extracted more relevant prosodic features from the Chinese text. After several times of feature selection, we get high accuracy and some prosodic features that can affect the nodding motion most. We use two machine learning algorithms to check whether the prediction model is good or not. We take Decision Tree algorithm as the baseline and use Neural Network algorithm to check the predicted effect. The Chinese corpus is sourced from Chinese "CCTV News". Objective and subjective experiments show that the proposed prediction model is good. |
Year | DOI | Venue |
---|---|---|
2016 | 10.1109/CIS.2016.101 | PROCEEDINGS OF 2016 12TH INTERNATIONAL CONFERENCE ON COMPUTATIONAL INTELLIGENCE AND SECURITY (CIS) |
Keywords | Field | DocType |
prediction model, Chinese talking avatar, prosodic features, Decision Tree, Neural Network | Computational intelligence,Feature selection,Computer science,Head movements,Motion generation,Speech recognition,Natural language processing,Artificial intelligence,Artificial neural network,Avatar,Decision tree learning | Conference |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yang Sun | 1 | 53 | 13.05 |
Yiwen Wang | 2 | 2 | 2.44 |