Sign in Subscribe
Machine learning

Artificial Intelligence Research Yields Breakthrough in Hemiplegic Gait Recognition

Strategic Insights

— 1 min read

Researchers at the Chengdu University of Technology have made significant strides in developing an artificial intelligence (AI) model that can accurately distinguish between hemiplegic gait and healthy gait. The study, published in the journal Computers, used a pressure sensor array to collect data from 19 hemiplegic patients and 29 healthy subjects. The AI model, known as the temporal-frequency domain interaction network (TFDI-Net), was trained to recognize patterns in the center of pressure (CoP) trajectory, a key indicator of gait abnormalities.

Key Takeaways:

  • The TFDI-Net model outperformed traditional machine learning methods, achieving improvements of 2.89% in recognition rate, 4.6% in F1-score, and 8.25% in recall.
  • The research was funded by the Chengdu Science And Technology Bureau and the Science And Technology Department of Sichuan Province.
  • The study collected CoP trajectory data from 48 subjects, including 19 hemiplegic patients and 29 healthy individuals.
  • The TFDI-Net model extracted frequency domain features from the CoP trajectory using fast Fourier transform (FFT) and interacted with time domain features to construct a discriminative joint representation.
  • The model was evaluated using five-fold cross-validation comparisons with traditional machine learning methods and deep learning methods.
  • Intra-fold data augmentation was performed to enhance the robustness of the model.

Statistics:

  • 19 hemiplegic patients and 29 healthy subjects were involved in the study.
  • The TFDI-Net model achieved a 2.89% improvement in recognition rate compared to traditional machine learning methods.
  • The model achieved a 4.6% improvement in F1-score and an 8.25% improvement in recall compared to traditional machine learning methods.
  • The study used a pressure sensor array to collect data from the subjects, with a total of 48 data points collected.

Sources:

  • Time-Frequency Feature Fusion Approach for Hemiplegic Gait Recognition. Computers, 2025, 14(8): 334.
  • Chengdu Science And Technology Bureau
  • Science And Technology Department of Sichuan Province
  • MDPI AG

Read more