A Convolutional Neural Network (CNN) is a deep learning architecture that applies learned spatial filters across input data through successive convolutional layers, enabling automatic extraction of hierarchical features without manual feature engineering. In the context of CSI-based Wi-Fi sensing, CNNs are particularly well-suited to processing structured, grid-like input representations — such as CSI amplitude images or colormap-transformed RGB tensors — allowing the model to capture local patterns in both frequency and temporal dimensions that are discriminative for tasks like occupancy detection and activity recognition. Key variants employed in this domain include standard 2D CNNs applied to spectrogram or image-formatted CSI inputs, as well as hybrid architectures that combine convolutional layers with recurrent or fully connected components to jointly model spatial and temporal dynamics across diverse traffic conditions, including ambient Wi-Fi streams.
Source Papers
- A low-cost automatic people-counting system at bus stops using Wi-Fi probe requests and deep learning ↗ — A low-cost automatic people-counting system at bus stops usi
- Channel State Information (CSI) Amplitude Coloring Scheme for Enhancing Accuracy of an Indoor Occupancy Detection System Using Wi-Fi Sensing ↗ — Channel State Information (CSI) Amplitude Coloring Scheme fo
- OPERAnet, a multimodal activity recognition dataset acquired from radio frequency and vision-based sensors ↗ — OPERAnet, a multimodal activity recognition dataset acquired
- Towards Energy Efficient Wireless Sensing by Leveraging Ambient Wi-Fi Traffic ↗ — Towards Energy Efficient Wireless Sensing by Leveraging Ambi