Elder falling detection method and system based on multiple classifier integration
A technology with multiple classifiers and detection methods, applied in the direction of instruments, alarms, etc., can solve the problem of low accuracy and achieve high accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment example 1
[0065] Implementation Example 1: Using a Support Vector Machine Classifier to Identify Falls
[0066] Support vector machine (Support Vector Machine, SVM) realizes automatic facial expression recognition. SVM is a classification method that has just been developed in recent years. It is based on the principle of structural risk minimization and has good generalization ability. Given a training sample set ,in is the input vector, For the corresponding category, SVM looks for the optimal boundary hyperplane that can correctly separate the two types of samples in the feature space. For a vector in the input space , if using Represents its corresponding eigenvector in the feature space, then the optimal boundary hyperplane is expressed as . The corresponding decision equation is . In any case, SVM does not require knowledge of the mapping Φ. Introduce kernel function , the dot product between vectors in the feature space can be expressed by the kernel function...
Embodiment example 2
[0081] Implementation Case 2: Using an Ensemble Classifier AdaBoost Identify falls
[0082] AdaBoost classifier is one of the top ten classification algorithms in data mining. It has the advantages of fast speed and simplicity. It does not need to adjust parameters except for the number of iterations, and does not require prior knowledge of weak classifiers. Given enough data and a moderately accurate weak classifier, it can promote the weak classifier to a strong classifier, thereby improving the recognition effect.
[0083] Different training sets in the AdaBoost classifier are achieved by adjusting the weights corresponding to each sample. At the beginning, the weight corresponding to each sample is the same, that is, for n samples, a weak classifier is trained under this sample distribution. For misclassified samples, increase their corresponding weights; and for correctly classified samples, reduce their weights, so that misclassified samples are highlighted, and a n...
Embodiment example 3
[0098] Implementation Example 3: Identifying Falls Using a Rotational Forest Classifier
[0099] Rotation forest is an ensemble learning method based on feature extraction method proposed by Juan J. Rodriguez et al. (Rodrignaz J J et al, Rotation forest: a new classifier ensemble method, TPAMI, 2006). The method first randomly divides the feature set into k subsets, where k is a parameter of the algorithm. Then apply principal component analysis (Principal component Analysis, PCA) on each partitioned subset. In order to preserve the information of the data in the method, all principal components will be retained. Using the PCA-based axis method has two purposes: to improve the performance of individual classifiers and to increase the diversity of all classifiers. The decision tree method is selected as the base classifier method, so the integrated method is called "rotation forest". The reason for choosing the decision tree as the base classifier is that it is sensitive to...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap