Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
You can not select more than 25 topics
Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
tal
a5282eb126
|
1 year ago | |
|---|---|---|
| config | 1 year ago | |
| LICENSE | 1 year ago | |
| README.md | 1 year ago | |
| model.py | 1 year ago | |
| run_iemocap-ours.py | 1 year ago | |
| run_ravdess-ours.py | 1 year ago | |
| utils.py | 1 year ago | |
README.md
Interspeech2021_CME
Model Code to Cross Modal Excitation
Some Processed data inputs can be download from: https://drive.google.com/drive/folders/1K9kqYMYGQQYq5RYKYp0IoGeu1rTfkqCx
Please note: some raw files are not shared due to the restriction of license.
If you need the raw files, please check the published data source's homepages:
IEMOCAP: https://sail.usc.edu/iemocap/index.html
RAVDESS: https://smartlaboratory.org/ravdess/
该算法提出一种新的多模态深度学习方法，可以识别现实生活说话中的细粒度情感。该算法包含了时间对齐的均值-最大值集合机制，以捕捉每个语篇中隐含的细节和细粒度的情绪。此外该算法还包含一个跨模态激活模块，可以实现对跨模态嵌入进行特定的调整，并通过它来对其他模态的潜在特征相对应的值进行校准。在 IEMOCAP 数据集上该算法WA 达到 72.7%，在 RAVDESS 数据集上 WA 达到 72%。
Python
