Related Experiment Video
Updated: Nov 3, 2025
06:37
Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
Published on: December 15, 2023
4.6K
Device-Free Human Activity Recognition with Low-Resolution Infrared Array Sensor Using Long Short-Term Memory Neural
Cunyi Yin1, Jing Chen1, Xiren Miao1
1College of Electrical Engineering and Automation, Fuzhou University, Fuzhou 350108, China.
Sensors (Basel, Switzerland)
|June 2, 2021
Summary
This study introduces a privacy-preserving human activity recognition system using low-resolution infrared sensors and deep learning. The system achieves high accuracy for daily activity classification, offering a cost-effective solution for smart environments.
Area of Science:
- Computer Science
- Artificial Intelligence
- Sensor Technology
Background:
- Human Activity Recognition (HAR) is crucial for IoT, smart homes, and healthcare.
- Existing HAR systems often face privacy concerns and high deployment costs.
- Device-free sensing offers a privacy-preserving alternative.
Purpose of the Study:
- To propose a novel low-resolution infrared array sensor-based HAR approach.
- To develop a privacy-preserving and cost-effective HAR solution.
- To leverage deep learning for accurate activity recognition.
Main Methods:
- Utilized an 8x8 pixel infrared array sensor for device-free sensing.
- Applied J-filter and Butterworth filters for infrared signal preprocessing to mitigate temperature variations.
- Employed Long Short-Term Memory (LSTM) networks for automatic feature extraction and model building.
- Designed a real-time HAR interface by embedding the LSTM model.
Main Results:
- Achieved a high recognition accuracy of 98.287% for typical daily activities.
- Demonstrated superior performance compared to existing machine learning methods.
- Validated the effectiveness of the proposed J-filter and Butterworth filter preprocessing techniques.
Conclusions:
- The proposed infrared sensor-based HAR system is a low-cost and effective solution.
- The approach successfully preserves user privacy while enabling accurate activity recognition.
- This technology holds significant promise for privacy-preserving applications in smart environments and healthcare.
