[论文笔记] Context-Aware Wireless-Based Cross-DomainGesture Recognition

最新推荐文章于 2023-06-21 11:12:38 发布

_孤鸿寄语_

最新推荐文章于 2023-06-21 11:12:38 发布

阅读量1.2k

点赞数

分类专栏：论文笔记文章标签：深度学习迁移学习

本文链接：https://blog.csdn.net/weixin_44563093/article/details/124207564

版权

论文笔记专栏收录该内容

17 篇文章 1 订阅

订阅专栏

论文链接： Context-Aware Wireless-Based Cross-Domain Gesture Recognition | IEEE Journals & Magazine | IEEE Xplore

Methodology [论文原图，侵删]

文章将一组environment-user pair看成一个domain，并没有利用Few shot learning

Fig 1. Train Model

Model input

A matrix with dimension $F \times M$ , where $F$ is the number of sampling points in the frequency domain, and $M$ is the transceiver link number. So the input DFS series are with dimension $T \times F \times M$ .

Feature Extractor

Fig 2. Architecture of feature extractor G

CNNs: extracts spatial features from each single DFS profile

GRUs: models the temporal dependencies of the whole sequence

为什么选择GRU：和LSTM相比paras更少！可有效防止过拟合。We choose GRUs because they are capable of learning longterm representations and they achieve comparable performance while involving fewer parameters compared to long short-term memory (LSTM).

希望提取的 $Z$ 能有一定的domain invariant 特性（可以通过后面的adversarial learning实现）

Domain Discriminator

文章针对每一个source domain $i$ 都设计了一个discriminator $D_{s_i}$ （总共N个），每个都有同样的结构：two fully connected layers where the first layer has leaky Relu activation while the second layer has sigmoid activation to get the output. 来自target domain的 $Z_t$ 会通过全部 D 但是来自 $i$ -th source domain的只会通过对应的discriminator $D_{s_i}$ 。

domain discriminators 的outputs 反映了target domain 和 source domain 之间的similarities. If the output is close to 0.5, it means that the domain discriminator is uncertain about whether the sample is from source or target.

Gesture Recognizer

Fig 3. Components of gesture recognizer $i$

同样针对每一个source domain $i$ 都设计了一个recognizer, 每个recognizer包含以下部分

1. Feature Disentanglement：之前提取的特征 $Z$ 可能依旧包含一些domain related features (location/orientation etc), 需要进一步的将其分离成 class-related features(gesture classification $CR_{s_i}$ ) 和 domain-related features (domain info regression $DR_{s_i}$ )

2. Gesture Classifier: 将获得的class-related features 进行后续分类

3. Domain Factor Regressor：将获得的domain-related features 输入并计算出和 ground-truth
location and orientation 的 regression loss

4. Feature Reconstructor：确保 extracted class-related features 和 domain-related features 可以恢复成 original feature （MSE loss）

Domain Attention and Class Operator [略]

_孤鸿寄语_

关注

0
点赞
踩
0

收藏

觉得还不错? 一键收藏
0
评论
[论文笔记] Context-Aware Wireless-Based Cross-DomainGesture Recognition

论文链接：Context-Aware Wireless-Based Cross-Domain Gesture Recognition | IEEE Journals & Magazine | IEEE XploreObjectiveMethodologyTrain ModelModel inputA matrix with dimension , where is the number of sampling points in the frequency domai...
复制链接

扫一扫