Small CSI Samples-Based Activity Recognition: A Deep Learning Approach Using Multidimensional Features

Tian, Yong; Li, Sirou; Chen, Chen; Zhang, Qiyue; Zhuang, Chuanzhen; Ding, Xuejun

doi:10.1155/2021/5632298

Cited by 6 publications

(3 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Experimental results show that ML-WIGR achieves crossdomain recognition accuracy of 87.8% to 94.8% with a small amount of training samples. Additionally, reference [24] proposes a bidirectional LSTM approach that utilizes multidimensional features and attention mechanisms. By expanding the training samples through tensor prediction algorithm, this model achieves an accuracy of 92%.…”

Section: Related Workmentioning

confidence: 99%

WiFi gesture recognition method based on data enhancement

Li,

Wang,

2024

Fourth International Conference on Machine Learning and Computer Application (ICMLCA 2023)

View full text Add to dashboard Cite

In recent years, human motion recognition based on wireless sensing has been widely used in smart home, virtual reality and other fields. Among them, human motion recognition based on WiFi channel state information (CSI) has received great attention. Although the current large-action recognition method based on CSI has good performance, the recognition accuracy is low when performing fine-grained human gesture recognition, which is due to the difficulty of collecting CSI human gesture perception dataset and the small number of high-quality datasets. Therefore, we convert CSI data into images and augment CSI images using super-resolution image generation methods. At the same time, we propose a SCBAM attention mechanism to further enhance the attention of the model to important features. In the experiment, we performed experimental analysis using the publicly available Widar3 dataset. The results show that the recognition accuracy of our mixed attention model in the same scene, across location, direction and scene reaches 99.73%, 97.41%, 94.57% and 95.98%, which is 0.07%, 0.29%, 0.05% and 0.99% higher than before data enhancement.

show abstract

Section: Related Workmentioning

confidence: 99%

WiFi gesture recognition method based on data enhancement

Li,

Wang,

2024

Fourth International Conference on Machine Learning and Computer Application (ICMLCA 2023)

View full text Add to dashboard Cite

show abstract

“…In addition, ResNet18 shows higher accuracy and less time consumption than other ResNet models. Moreover, MF-ABLSTM [36] leverages attention-based long short-term memory neural network and time-frequency domain features for small CSI sample-based activity recognition, achieving 92% with 490 training and testing samples after 200 iterations. Due to the proposed method being able to train very deep neural networks, it avoids the problem of gradient vanishing and improves the model's expressive power and performance.…”

Section: Experimental Validationmentioning

confidence: 99%

Transfer-Learning-Based Human Activity Recognition Using Antenna Array

Ye,

Wu,

Cai

et al. 2024

Remote Sensing

View full text Add to dashboard Cite

Due to its low cost and privacy protection, Channel-State-Information (CSI)-based activity detection has gained interest recently. However, to achieve high accuracy, which is challenging in practice, a significant number of training samples are required. To address the issues of the small sample size and cross-scenario in neural network training, this paper proposes a WiFi human activity-recognition system based on transfer learning using an antenna array: Wi-AR. First, the Intel5300 network card collects CSI signal measurements through an antenna array and processes them with a low-pass filter to reduce noise. Then, a threshold-based sliding window method is applied to extract the signal of independent activities, which is further transformed into time–frequency diagrams. Finally, the produced diagrams are used as input to a pretrained ResNet18 to recognize human activities. The proposed Wi-AR was evaluated using a dataset collected in three different room layouts. The testing results showed that the suggested Wi-AR recognizes human activities with a consistent accuracy of about 94%, outperforming the other conventional convolutional neural network approach.

show abstract

“…Since CSI can be essentially retrieved as time-series data, the paradigm of Deep Learning (DL) represents a straightforward approach to the data-driven modeling of HAR systems to be adopted in this context. Namely, Recurrent Neural Networks (RNNs) are one of the most reliable tools to develop HAR using Wi-Fi; in fact, many accurate solutions for time series analysis are based on Deep RNNs [1], where information flows through several neural layers organized in sequential stacks or acyclic graphs. However, while being quite accurate and precise, Deep RNNs are computationally expensive, and they require extensive time and computing resources.…”

Section: Introductionmentioning

confidence: 99%