A Combined Method Based on SVM and Online Learning with HOG for Hand Shape Recognition

2012 ◽  
Vol 16 (6) ◽  
pp. 687-695 ◽  
Author(s):  
Kazutaka Shimada ◽  
◽  
Ryosuke Muto ◽  
Tsutomu Endo
Keyword(s):  
Online Learning ◽  
Learning Algorithm ◽  
Shape Recognition ◽  
Input Image ◽  
Training Data ◽  
Support Vector ◽  
Combined Method ◽  
Hand Shape ◽  
Vector Machines ◽  
Online Learning Algorithm

In this paper, we propose a combined method for hand shape recognition. It consists of Support Vector Machines (SVMs) and an online learning algorithm based on the perceptron. We apply HOG features to each method. First, our method estimates the hand shape of an input image by using SVMs. Here, an online learning method with the perceptron uses an input image as new training data if the image is effective in relearning in the recognition process. Next, we select a final hand shape from the outputs of SVMs and perceptrons by using the score from SVMs. The combined method with the online perceptron is robust against unknown users because it contains a relearning process for the current user. Therefore applying the online perceptron leads to an improvement in accuracy. We compare the combined method with a method that uses only SVMs. Experimental results show the effectiveness of the proposed method.

An Online Prediction Model for BFG Output in Steel Industry

2012 ◽  
Vol 542-543 ◽  
pp. 507-512 ◽  
Author(s):  
Xiaoping Zhang ◽  
Jun Zhao
Keyword(s):  
Prediction Model ◽  
Steel Industry ◽  
Learning Algorithm ◽  
Difficult Problem ◽  
Parameters Optimization ◽  
Training Data ◽  
Support Vector ◽  
Mode Decomposition ◽  
Online Learning Algorithm ◽  
Gas Balance

The output prediction of blast furnace gas (BFG), influenced by many complex production factors, is a very important and difficult problem concerning the byproduct gas balance in steel industry. A new online least squares support vector machine (LSSVM) prediction model is proposed in this paper, in which the training data is filtered by an improved empirical mode decomposition threshold filtering (IEMDTF). The model is solved and optimized by an online learning algorithm and an online bayesian parameters optimization, respectively. The experimental results using practical BFG output data from BaoSteel Co. Ltd., China show the proposed model is effective and enable to offer reasonable gas balance scheduling for operators.

Online Learning Algorithm of Gaussian Process Based on Adaptive Natural Gradient for Regression

2010 ◽  
Vol 139-141 ◽  
pp. 1847-1851 ◽  
Author(s):  
Qian Qian Shen ◽  
Zong Hai Sun
Keyword(s):  
Online Learning ◽  
Gaussian Process ◽  
Learning Algorithm ◽  
System Modeling ◽  
Training Data ◽  
Data Sets ◽  
Continuous Stirred Tank Reactor ◽  
Natural Gradient ◽  
Nonlinear System Modeling ◽  
Online Learning Algorithm

Gaussian Process (GP) is a new learning method on nonlinear system modeling. The most common way of model training is conjugate gradient method, but this method should compute Heisenberg matrix which needs much computing resource. It is not a suitable training method for online learning algorithm. There is one online learning algorithm of GP which is named sparse online GP now. This algorithm has constraint to the training data sets. In order to satisfy the real-time modeling without the limit of the training data sets, an online algorithm of GP based on adaptive natural gradient (ANG) is proposed in this paper. The algorithm is applied in Continuous Stirred Tank Reactor (CSTR) modeling and the sparse online GP is also applied in CSTR modeling for comparison. Obtained from the simulation results, the algorithm is effective and has higher Accuracy compared with the sparse online GP algorithm.

A Fast Unsupervised Online Learning Algorithm to Detect Structural Damage in Time-Varying Environments

2021 ◽  
Author(s):  
Karthik Gopalakrishnan ◽  
V. John Mathews
Keyword(s):  
Online Learning ◽  
Damage Detection ◽  
Structural Damage ◽  
Learning Algorithm ◽  
Operating Conditions ◽  
Training Data ◽  
Support Vector ◽  
Support Vector Data Description ◽  
Time Varying ◽  
Varying Environments

Abstract Machine learning based health monitoring techniques for damage detection have been widely studied. Most such approaches suffer from two main problems, time-varying environmental and operating conditions, and the difficulty in acquiring training data from damaged structures. Recently, our group presented an unsupervised learning algorithm using support vector data description (SVDD) and an autoencoder to detect damage in time-varying environments without training on data from damaged structures. Though the preliminary experiments produced promising results, the algorithm was computationally expensive. This paper presents an iterative algorithm that learns the state of a structure in time-varying environments online in a computationally efficient manner. This algorithm combines the fast, incremental SVDD (FISVDD) algorithm with signal features based on wavelet packet decomposition (WPD) to create a method that is efficient and provides more accurate detection of smaller damage than the autoencoder-based method. The use of FISVDD has created the possibility of online learning and adaptive damage detection in time-varying environmental and operating conditions (EOC). The WPD-based features also have the potential to provide explainability for the learning algorithm.

Sign in / Sign up

Export Citation Format

Share Document