Description

A spectrogram is the magnitude of an STFT, plotted as a 2D image with time on one axis and frequency on the other. CSI / radar spectrograms are the canonical input representation for CNN-based HAR, gesture, gait, and fall pipelines because human motion produces visually distinctive Doppler patterns at familiar locomotion frequencies.

When it's used

  • Input to CNN backbones for HAR / gesture / gait
  • Visual analysis of micro-Doppler signatures
  • Doppler-bin extraction for crowd flow

Limitations

  • Throws away phase information
  • Time-frequency resolution trade-off in window choice
  • Hardware artefacts (CFO drift) can dominate the low-frequency band

Source Papers

  • wang2016_6482 — gait spectrograms
  • ullmann2023_0ac3 — radar spectrogram HAR
  • li2021_1875 — spectrogram in passive WiFi radar
  • ren2023_8cfe — spectrogram processing in radar pipelines

11 vault papers use this method

Titles and DOIs only — no abstracts, no analyses.

  • Understanding and Modeling of WiFi Signal Based Human Activity Recognition 2015 DOI ↗
  • Gait recognition using wifi signals 2016 DOI ↗
  • Passive WiFi Radar for Human Sensing Using a Stand-Alone Access Point 2021 DOI ↗
  • On CSI and Passive Wi-Fi Radar for Opportunistic Physical Activity Recognition 2022 DOI ↗
  • Efficient machine learning for Wi-Fi CSI-based human activity recognition using fast Monte Carlo based feature extraction 2026 DOI ↗
  • WiFi-Based Human Sensing With Deep Learning: Recent Advances, Challenges, and Opportunities 2024 DOI ↗
  • OPERAnet, a multimodal activity recognition dataset acquired from radio frequency and vision-based sensors 2022 DOI ↗
  • OPERAnet, a multimodal activity recognition dataset acquired from radio frequency and vision-based sensors 2022 DOI ↗
  • CSI crowd-counting: An experimental study using Machine Learning and Deep Learning Algorithms 2023 DOI ↗
  • CSI crowd-counting: An experimental study using Machine Learning and Deep Learning Algorithms 2023 DOI ↗
  • CSI crowd-counting: An experimental study using Machine Learning and Deep Learning Algorithms 2023 DOI ↗