Face recognition (FR), the process of identifying people through facial images, has numerous practical applications in the area of biometrics, information security, access control, law enforcement, smart cards and surveillance system. Convolutional Neural Networks (CovNets), a type of deep networks has been proved to be successful for FR. For real-time systems, some preprocessing steps like sampling needs to be done before using to CovNets. But then also complete images (all the pixel values) are passed as input to CovNets and all the steps (feature selection, feature extraction, training) are performed by the network. This is the reason that implementing CovNets are sometimes complex and time consuming. CovNets are at the nascent stage and the accuracies obtained are very high, so they have a long way to go. The paper proposes a new way of using a deep neural network (another type of deep network) for face recognition. In this approach, instead of providing raw pixel values as input, only the extracted facial features are provided. This lowers the complexity of while providing the accuracy of 97.05% on Yale faces dataset.

This paper takes Twitter as the framework and intended to propose an optimum approach for classification of Twitter data on the basis of the contextual and lexical aspect of tweets. It is a dire need to have optimum strategies for offensive content detection on social media because it is one of the most primary modes of communication, and any kind of offensive content transmitted through it may harness its benefits and give rise to various cyber-crimes such as cyber-bullying and even all content posted during the large even on twitter is not trustworthy. In this research work, various facets of assessing the credibility of user generated content on Twitter has been described, and a novel real-time system to assess the credibility of tweets has been proposed by assigning a score or rating to content on Twitter to indicate its trustworthiness. A comparative study of various classifying techniques in a manner to support scalability has been done and a new solution to the limitations present in already existing techniques has been explored.