Pattern recognition

This paper presents a deep learning-based sensing fusion system to detect and recognize actions of interest from continuous action streams, which contain actions of interest occurring continuously and randomly among arbitrary actions of non-interest. The sensors used in the fusion system consist of a depth camera and a wearable inertial sensor. A convolutional neural network is utilized for depth images obtained from the depth sensor, and a combination of convolutional neural network and long short-term memory network is utilized for inertial signals obtained from the inertial sensor. Each sensing modality first performs segmentation of all actions and then detection of actions of interest for a particular application. A decision-level fusion of the two sensing modalities is carried out to achieve the recognition of the detected actions of interest. The developed fusion system is examined for two applications: one involving transition movements for home healthcare monitoring and the other involving smart TV hand gestures. The results obtained show the effectiveness of the developed fusion system in dealing with realistic continuous action streams.

Action Detection and Recognition in Continuous Action Streams by Deep Learning-Based Sensing Fusion

Pattern analysis often requires a pre-processing stage for extracting or selecting features in order to help the classification, prediction, or clustering stage discriminate or represent the data in a better way. The reason for this requirement is that the raw data are complex and difficult to process without extracting or selecting appropriate features beforehand. This paper reviews theory and motivation of different common methods of feature selection and extraction and introduces some of their applications. Some numerical implementations are also shown for these methods. Finally, the methods in feature selection and extraction are compared.

Feature Selection and Feature Extraction in Pattern Analysis: A Literature Review.

This tutorial explains Linear Discriminant Analysis (LDA) and Quadratic Discriminant Analysis (QDA) as two fundamental classification methods in statistical and probabilistic learning. We start with the optimization of decision boundary on which the posteriors are equal. Then, LDA and QDA are derived for binary and multiple classes. The estimation of parameters in LDA and QDA are also covered. Then, we explain how LDA and QDA are related to metric learning, kernel principal component analysis, Mahalanobis distance, logistic regression, Bayes optimal classifier, Gaussian naive Bayes, and likelihood ratio test. We also prove that LDA and Fisher discriminant analysis are equivalent. We finally clarify some of the theoretical concepts with simulations we provide.

Linear and Quadratic Discriminant Analysis: Tutorial.

As depression becomes more commonplace in society, the timely and effective detection of the signs of depression for its prevention and early treatment becomes more important. Gait analysis can provide a contactless and low-cost method for depression diagnosis. In this study, we propose a novel gait assessment framework to implement non-intrusive, real-time and automatic depression detection using Kinect, an inexpensive and portable depth sensor. We focus on extracting a novel time-domain and frequency-domain feature (TF-feature) and a spatial geometric feature (SG-feature), and investigating the effectiveness of fused features in detecting depression for the non-contact gait data. A pseudo-velocity model is firstly built to analyze the gait abnormalities of individuals with depression in the time domain. Subsequently, we perform the power spectral density (PSD) analysis on the model to extract the TF-feature. Then, the covariance matrices and the symmetric Stein divergence (S-divergence) are leveraged to obtain the SG-feature, which is fused with TF-feature to form new features for classification. The experimental results on 95 subjects (43 scored-depressed and 52 non-depressed individuals) show that the proposed method achieves a good classification accuracy of 93.75%, has superior performance compared to several other methods, and significantly alleviates the impact of individual differences. These results indicate the efficacy and robustness of the proposed framework for depression detection.

A Gait Assessment Framework for Depression Detection Using Kinect Sensors

This paper proposes a new method for view-invariant action recognition that utilizes the temporal position of skeletal joints obtained by Kinect sensor. In this method, the actions are represented as sequences of several pre-defined poses. After pre-processing, which includes skeleton alignment and scaling, the appropriate feature vectors are obtained for recognizing and discriminating the pose of every frame by the proposed Fisherposes method. The proposed regularized Mahalanobis distance metric is used in order to recognize both the involuntary and highly made-up actions at the same time. Hidden Markov model (HMM) is then used to classify the action related to an input sequence of poses. For taking into account the motion in the actions which are not separable by solely their temporal poses, histograms of trajectories are also proposed. The proposed action recognition method is capable of recognizing both the voluntary and involuntary actions, as well as pose-based and trajectory-based ones with a high accuracy rate. The effectiveness of the proposed method is experimented on three publicly available data sets, TST fall detection, UTKinect, and UCFKinect data sets.

Fisherposes for Human Action Recognition Using Kinect Sensor Data

Human action recognition has been one of the most active fields of research in computer vision for last years. Two dimensional action recognition methods are facing serious challenges such as occlusion and missing the third dimension of data. Development of depth sensors has made it feasible to track positions of human body joints over time. This paper proposes a novel method of action recognition which uses temporal 3D skeletal Kinect data. This method introduces the definition of body states and then every action is modeled as a sequence of these states. The learning stage uses Fisher Linear Discriminant Analysis (LDA) to construct discriminant feature space for discriminating the body states. Moreover, this paper suggests the use of the Mahalonobis distance as an appropriate distance metric for the classification of the states of involuntary actions. Hidden Markov Model (HMM) is then used to model the temporal transition between the body states in each action. According to the results, this method significantly outperforms other popular methods, with recognition rate of 88.64% for eight different actions and up to 96.18% for classifying fall actions.

Recognizing Involuntary Actions from 3D Skeleton Data Using Body States

http://scientiairanica.sharif.edu/article_20446_3aef8aadd7a2890184233d8738c269b7.pdf

Performance of Friction Stir Welding (FSW) as a solid-state process is approved in several engineering applications, especially aluminum industries. Identification of mechanical behavior of the associated welded zone is necessary due to these extensive applications of FSW. In this study, considering the effect of rotational and forward speed of welding tool on the mechanical properties of welded region, a hybrid optimization method based on combination of Genetic Algorithm (GA) and Response Surface Method (RSM) named here as GA-RSM is proposed to achieve maximum tensile and ultimate strength. The results of GA-RSM are validated by per-forming necessary experimental tests on two wide-used 2024 and 5050 aluminum alloys. The results show that GA-RSM could be an effective approach to achieve optimized process for FSW with minimum cost.

/pdf/development-of-an-optimal-process-for-friction-stir-welding-266g0pncsc.pdf

Mozhgan Mokari

Papers

Fisherposes for Human Action Recognition Using Kinect Sensor Data

Recognizing Involuntary Actions from 3D Skeleton Data Using Body States

Recognizing Involuntary Actions from 3D Skeleton Data Using Body States

Development of an optimal process for friction stir welding based on GA-RSM hybrid algorithm