| 余金锁,卢先领.基于分割注意力的特征融合 CNN-Bi-LSTM
人体行为识别算法[J].电子测量与仪器学报,2022,36(2):89-95 |
| 基于分割注意力的特征融合 CNN-Bi-LSTM
人体行为识别算法 |
| Human action recognition algorithm of feature fusionCNN-Bi-LSTM based on split-attention |
| |
| DOI： |
| 中文关键词: 行为识别 分割注意力 特征融合 双向长短时记忆网络 |
| 英文关键词:action recognition split-attention feature fusion BI-LSTM |
| 基金项目:国家自然科学基金（61773181）项目资助 |
|
|
|
|
| 摘要点击次数: 299 |
| 全文下载次数: 778 |
| 中文摘要: |
| 针对传统人体行为识别算法不能有效抑制空间背景信息,网络间缺乏信息交互,以及无法对全局时间相关性进行建模
的问题,提出一种基于分割注意力的特征融合卷积神经网络-双向长短时记忆网络(CNN-Bi-LSTM)人体行为识别算法。 首先以
一定采样率采样 30 帧图像,通过分割注意力网络提取图像的深度特征,并引入特征融合机制增强不同卷积层间的信息交互;然
后将深度特征输入到 Bi-LSTM 网络对人体动作的长时时间信息建模,最后使用 Softmax 分类器对识别结果进行分类。 相较于传
统双流卷积网络,该算法在 UCF101 和 HMDB51 数据集上的准确率分别提高了 6. 6%和 10. 2%,有效提高了识别准确率。 |
| 英文摘要: |
| Aiming at the problems that traditional human action recognition algorithms cannot effectively suppress spatial background
information, the lack of information interaction between networks, and the inability to model global temporal correlation, a human action
recognition algorithm of feature fusion Bi-LSTM based on segmentation attention is proposed. First, 30 frames of images are sampled at a
certain sampling rate, extract the depth features of the images by split-attention network, and introduce a feature fusion mechanism to
enhance the information interaction between different convolutional layers. Then input the depth features into the Bi-LSTM network to
model the long-term information of human actions, and finally use the Softmax classifier to classify the recognition results. Compared with
the traditional two-stream convolutional network, the accuracy of this algorithm on the UCF101 and HMDB51 datasets is increased by
6. 6% and 10. 2%, respectively, which effectively improves the recognition accuracy. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
