In image processing is to identify the human faces in difficult to recognizing image analysis which has each day most applications. The main idea in the building of the detector is a learning classification built on ada-boost. The families of simple classifiers contain simple rectangular wavelets which are reminiscent of the Haar basis. Their ease and a new image representation called Integral Image allow a very quick computing of these Haar-like features. An assembly in cascade is introduced in order to reject quickly the easy to classify background regions and focus on the inflexible to classify windows. The structure of the SVM classifier allows a real-time implementation of the indicator. Some results on real world examples are presented. The detector yields good detection rates with frontal faces then the process can be easily adapted to other object detection tasks by changing the contents of the training dataset.
Wireless Mesh Networks (WMNs) are a promising technology that could revolutionize the way wireless network access is provided. We allocate the available bandwidth in wireless mesh network to reducing the energy consumption of a user using the AODV protocol. the AODV protocol to reducing energy consumption using the improving parameters are throughput, end to end delay, network overhead, energy spent, packet delivery ratio. The bandwidth allocation using greedy algorithm and gateway load balancing routing protocol. In this paper Genetic Algorithm used for providing solution in optimization problem.Topology of wireless mesh networks with more no of nodes is routed using genetic algorithm with proposed approaches and parameters such as end to end delay, throughput, energy spent, packet delivery ratio.finally comparison done between greedy algorithm and genetic algorithm.
The need for privacy preserving is sharing of sensitive information occurs in many different ways. In order to maintain privacy in database, the confidential data should be protected in the form of modifying the sensitive data items. Protecting sensitive data is an important issue in the government, public database. It used protected the sensitive numerical data item in the form of modifying the original data item using the proposed techniques. There are various anonymization technique provides privacy protection which can be used such as data encryption, Randomization and k-anonymity. The existing system uses commutative encryption scheme to improve data privacy of the database and provides security of data by using AES algorithm. To enhancing other cryptography techniques (such as RSA) are used in the user database for secure their database from unauthorized user. The multiple users have to access the data from a database with the permission of administrator.
To diminish the behavioral variability of mouse dynamics, the machine learning algorithm was proposed. Mouse dynamics is the process of identifying the user based on their mouse operating behavior. The dataset includes co-ordinates values, time stamp value and mouse operation. From this dataset, the schematic features, holistic features and motor-skill features like average speed, average distance, mean, standard deviation and mouse silence ratio, velocity, slope angle, curvature were extracted to obtain feature vector. The obtained feature vector can be applied to the dimensionality reduction based approach, diffusion map to reduce the dimension of the feature vector that compared with ISOMAP (Isometric Feature Mapping). Without dimensionality reduction based method the classification process was difficult. The machine learning algorithm i.e.) hop field network to be used to identify whether the given input sample was authenticated user (or) unauthenticated.