HUMAN MOTION ANALYSIS

Abstract

Human motion analysis in the video has its vast application. The recognition of the human action is the most prominent application of human motion analysis. In this research we analyzed different methodologies for modeling human action. We also discussed challenges and methodologies which are used to handle them. These methodologies are divided into two categories. One is global feature descriptor and other is local feature descriptors. The disadvantage of the global feature descriptor is that they can only give the shape information but fails to give motion information. The local feature descriptors are used to find out the motion information of the action video. The disadvantage is that it cannot give the shape or structure information of the action video. The hybrid descriptors are used to solve these problems but these descriptors suffer from high dimensionality features. In this research we proposed new feature descriptors which are capable to deal with these issues in the following manner: 1) We proposed a new local descriptor evaluated from the Finite Element Analysis for human action recognition. This local descriptor represents the distinctive human poses in the form of the stiffness matrix. This stiffness matrix gives the information of motion as well as shape change of the human body while performing an action. Initially, the human body is represented in the silhouette form. Most prominent points of the silhouette are then selected. This silhouette is discretized into several finite small triangle faces (elements) where the prominent points of the boundaries are the vertices of the triangles. The stiffness matrix of each triangle is then calculated. The feature vector representing the action video iii frame is constructed by combining all stiffness matrices of all possible triangles. These feature vectors are given to the Radial Basis Function-Support Vector Machine (RBF-SVM) classifier. The proposed method shows its superiority over other existing state-of-the-art methods on the challenging datasets Weizmann, KTH, Ballet, and IXMAS. 2) Background cluttering, appearance change due to variation in viewpoint and occlusion are the prominent hurdles that can reduce the recognition rate significantly. Methodologies based on Bag-of-visual-words are very popular because they do not require accurate background subtraction. But the main disadvantage with these methods is that they do not retain the geometrical structural information of the clusters that they form. As a result, they show intra-class mismatching. Furthermore, these methods are very sensitive to noise. Addition of noise in the cluster also results in the misclassification of the action. To overcome these problems we proposed a new approach based on modified Bag-of-visual-word. Proposed methodology retains the geometrical structural information of the cluster based on the calculation of contextual distance among the points of the cluster. Normally contextual distance based on Euclidean measure cannot deal with the noise but in the proposed methodology contextual distance is calculated on the basis of a difference between the contributions of cluster points to maintain its geometrical structure. Later directed graphs of all clusters are formed and these directed graphs are described by the Laplacian. Then the feature vectors representing Laplacian are fed to the Radial Basis Function based Support Vector Machine (RBF-SVM) classifier. iv 3) We also proposed a new feature descriptor to detect abnormality in a video captured for surveillance applications in real-time and also overcome the problem of the curse of dimensionality. To extract features related to any change in the video, non linear Gaussian fuzzy lattice functions have been applied on each frame of the video which results in the formation of fuzzy lattices. These fuzzy lattices have been expressed in the form of Schrödinger equation to find the kinetic energy involved corresponding to any change in the video. A number of the fuzzy lattice has been used as a dimension of the feature. It reduces the dimensionality significantly as compared to other state-of-the-art methods. Finally, the kinetic energy parameter is classified into normal and abnormal activities with the help of SVM-based classifier.