Related Experiment Video
Updated: Oct 27, 2025
08:05
Design and Analysis for Fall Detection System Simplification
Published on: April 6, 2020
10.9K
The Performance of Post-Fall Detection Using the Cross-Dataset: Feature Vectors, Classifiers and Processing
Bummo Koo1, Jongman Kim1, Yejin Nam1
1Department of Biomedical Engineering, Yonsei University, Wonju 26493, Korea.
Sensors (Basel, Switzerland)
|July 24, 2021
Summary
This study evaluated post-fall detection algorithms using machine learning. Results show that diverse training data improves accuracy, while specific data processing techniques impact classifier performance for fall detection systems.
Area of Science:
- Biomedical Engineering
- Machine Learning
- Wearable Technology
Background:
- Fall detection systems are crucial for elderly care and patient monitoring.
- Accurate algorithms are needed to distinguish falls from normal activities.
- Wearable inertial measurement units (IMUs) offer a promising approach for real-time fall detection.
Purpose of the Study:
- To evaluate the performance of different algorithms for post-fall detection.
- To assess the impact of feature vectors, classifiers, and data processing on detection accuracy.
- To investigate the generalizability of fall detection models across different datasets.
Main Methods:
- Collected three-axis acceleration and angular velocity data from 30 healthy males using IMUs attached to the anterior superior iliac spines (ASIS).
- Evaluated Artificial Neural Networks (ANN) and Support Vector Machines (SVM) classifiers using time-series and discrete feature vectors.
- Tested algorithms using internal (lab) and external (SisFall public) datasets under various processing conditions: normalization, equalization, increased training data, and external data training.
Main Results:
- ANN and SVM classifiers demonstrated suitability for time-series and discrete data, respectively.
- Classification performance decreased significantly when tested on untrained motions from an external dataset, highlighting the need for raw data-specific feature vectors.
- Data normalization improved SVM effectiveness but reduced ANN performance; equalization enhanced sensitivity without improving overall performance.
- Increasing the volume of training data positively impacted classification accuracy.
Conclusions:
- Machine learning models for fall detection are vulnerable to untrained motions, necessitating diverse movement data for robust training.
- Feature vector selection and data processing techniques critically influence the performance of fall detection algorithms.
- Cross-dataset validation is essential to assess the generalizability and limitations of developed fall detection systems.
Related Concept Videos
01:22Force Classification
1.9K
Forces play a crucial role in the study of physics and engineering. They are essential in describing the motion, behavior, and equilibrium of objects in the physical world. Forces can be classified based on their origin, type, and direction of action.
Contact and non-contact forces are two of the most widely used categories of forces. As the name suggests, contact forces require physical contact between two objects to act upon each other. Examples of contact forces include frictional,...
Contact and non-contact forces are two of the most widely used categories of forces. As the name suggests, contact forces require physical contact between two objects to act upon each other. Examples of contact forces include frictional,...
1.9K
01:30Classification of Signals
1.0K
In signal processing, signals are classified based on various characteristics: continuous-time versus discrete-time, periodic versus aperiodic, analog versus digital, and causal versus noncausal. Each category highlights distinct properties crucial for understanding and manipulating signals.
A continuous-time signal holds a value at every instant in time, representing information seamlessly. In contrast, a discrete-time signal holds values only at specific moments, often denoted as x(n), where...
A continuous-time signal holds a value at every instant in time, representing information seamlessly. In contrast, a discrete-time signal holds values only at specific moments, often denoted as x(n), where...
1.0K
01:29Aggregates Classification
423
Aggregate classification is generally based on its size, petrographic characteristics, weight, and source. Size classification ranges from coarse to fine aggregates, defined by the size of the particles. Coarse aggregates are particles that do not pass through ASTM sieve No. 4, and aggregates that pass through the sieve are fine aggregates.
Petrographic classification groups aggregates based on common mineralogical characteristics. Some of the common mineral groups found in aggregates are...
Petrographic classification groups aggregates based on common mineralogical characteristics. Some of the common mineral groups found in aggregates are...
423
01:26Classification of Systems-I
371
Linearity is a system property characterized by a direct input-output relationship, combining homogeneity and additivity.
Homogeneity dictates that if an input x(t) is multiplied by a constant c, the output y(t) is multiplied by the same constant. Mathematically, this is expressed as:
Homogeneity dictates that if an input x(t) is multiplied by a constant c, the output y(t) is multiplied by the same constant. Mathematically, this is expressed as:
371
01:31Classification of Systems-II
268
Continuous-time systems have continuous input and output signals, with time measured continuously. These systems are generally defined by differential or algebraic equations. For instance, in an RC circuit, the relationship between input and output voltage is expressed through a differential equation derived from Ohm's law and the capacitor relation,
268