A stochastic neural model for fast classification of binary images
IJCNN'09 Proceedings of the 2009 international joint conference on Neural Networks
Understanding transit scenes: a survey on human behavior-recognition algorithms
IEEE Transactions on Intelligent Transportation Systems
| Hi-index | 0.00 |
In this paper, we present human mistrustful motion detection & classification using Hu Moment Invariants feature descriptions. A new method for recognition & classification that is Moment Invariant based Classifier (MIBC) has been proposed. The basis of the MIBC is the different seven φ values of Hu Moment Invariants itself and the Euclidean Distance measure between these φ values of all image frames of each type of motion. These values of Moment Invariants and Euclidean Distances are compared with other values and Euclidean Distances of all image frames of different or same type of motion. The performance of MIBC is evaluated with other types of methodologies & classifiers like Mahalanobis Distance (MD) classifier, Linear Bayes Gaussian (LBG) classifier, Quadratic Bayes Gaussian (QBG) classifier and Fuzzy K-Nearest Neighbor (FKNN) classifier using temporal template motion detection technique. The performance evaluation is done in the context of accuracy, time and speed. Experiments are conducted on five types of suspicious motions: Bending down, Gun Shot, Jumping up, Kicking front and Punching forward.