Multiclass object recognition using smart phone and cloud computing for augmented reality and video surveillance applications

This paper presents multiclass object classification and recognition using smartphone and cloud computing (client server) technology. Smart phone camera is used as image acquisition device. Smartphone is working as a client and high speed computer act as a server. Our system is a feature based novel approach that requires huge computing power and stand-alone smart phone is not capable for performing the whole task. So we have used the smart phone as image acquisition and rendering device, it is also worked as a client and high speed computer server is used as a major computing unit like a cloud. We have adapted the bag of words approach for training features of multiclass objects with the usage of visual codebooks which are having significant applications in the natural language processing. Our work is mainly focused on classification and recognition of multiclass natural objects which can be utilized either for augmented reality and also video surveillance applications. We use Scale Invariant Feature Transforms (SIFT) for feature extraction. We form visual codebook from the high dimensional feature vectors using clustering algorithm and classify and recognize using naïve Bayes classifier.