Speedup Robust Feature and Support Vector Machine based Eco_Centric Activity Recognition
In this System, the problem of recognizing human activities from video sequences are discussed. The need for
such systems is increasing every day, with the number of (hundreds or thousands) of surveillance cameras deployed in public
spaces. This large number of cameras calls for systems able to detect, categorize and recognize human activity, requesting
human attention only when necessary. Our work is focused on three fundamental issues: (1) the design of a classifier and
data modeling for activity recognition; (2) how to perform feature selection and (3) how to define the structure of a classifier.
The recognition process is performed by using Naive Bayes, AdaBoost, KNN and neural network classifiers in existing
system. We use of a SVM classifier, and model the likelihood functions as Gaussian mixtures, adequate to cope with
complex data distributions, that are learned automatically. As for feature selection, we use SURF (Speedup Robust Features)
key points, to extract the key points from the images. SVM is applied on SURF features to provide better class separation.
Experimental results show that the proposed method meets out the requirements of the Eco centric activity system such as
imperceptibility, capacity and robustness. The output of the proposed method is superior to the existing methods.
Keywords - Speedup Robust Feature, Support Vector Machine, Activity Recognition.