Recent Advances in Computer Science and Communications - Volume 13, Issue 2, 2020

Background: Enabling fast and reliable lane-lines detection and tracking for advanced driving assistance systems and self-driving cars. Methods: The proposed technique is mainly a pipeline of computer vision algorithms that augment each other and take in raw RGB images to produce the required lane-line segments that represent the boundary of the road for the car. The main emphasis of the proposed technique in on simplicity and fast computation capability so that it can be embedded in affordable CPUs that are employed by ADAS systems. Results: Each used algorithm is described in details, implemented and its performance is evaluated using actual road images and videos captured by the front mounted camera of the car. The whole pipeline performance is also tested and evaluated on real videos. Conclusion: The evaluation of the proposed technique shows that it reliably detects and tracks road boundaries under various conditions.

Objective: Humans with their developed senses can easily ascertain a person’s gender just by listening to a few uttered words and it does not take any conscious additional effort to do so, however, a machine cannot do the same unless trained. This research proposes a real time system to identify a person’s gender from their voice. Method: Features are extracted from the dataset and checked for outliers. Then a baseline classifier is constructed to measure performance of the different models. Next the dataset is prepared for training and five machine learning models, Decision Tree Classifier, Random Forest Classifier, K Nearest Neighbours, Support Vector Machine and Gaussian Naive Bayes Classifier are applied. Finally, real time prediction is done by taking speech input and analysing it against the trained model, after input of speech the gender along with accuracy of prediction is displayed within 1.37s. Results: A maximum accuracy score of 88.19% is obtained using SVM. Additionally, the juxtaposition of the feature importance graph highlights the two most important features which fuel this classification. A combination of these features is then studied to design a less complex system and it is observed that using just MFCCs and Chroma Vector a near optimal accuracy score of 87.78% is obtained. Conclusion: Identification of gender prior to applying speech recognition and emotion recognition algorithm can help in reduction of the search space. Further, using only MFCC and Chroma Vector can make the system memory efficient and yet provide near optimal accuracy. The system can be used as an authentication mechanism and can be installed in public places.

Background: Big data analytics is the process of utilizing a collection of data accompanied on the internet to store and retrieve anywhere and at any time. Big data is not simply a data but it involves the data generated by variety of gadgets or devices or applications. Objective: When massive volume of data is stored, there is a possibility for malevolent attacks on the searching data are stored in the server because of under privileged privacy preserving approaches. These traditional methods result in many drawbacks due to various attacks on sensitive information. Hence, to enhance the privacy preserving for sensitive information stored in the database, the proposed method makes use of efficient methods. Methods: In this manuscript, an optimal privacy preserving over big data using Hadoop and mapreduce framework is proposed. Initially, the input data is grouped by modified fuzzy c means clustering algorithm. Then we are performing a map reduce framework. And then the clustered data is fed to the mapper; in mapper the privacy of input data is done by convolution process. To validate the privacy of input data the recommended technique utilizes the optimal artificial neural network. Here, oppositional fruit fly algorithm is used to enhancing the neural networks. Results: The routine of the suggested system is assessed by means of clustering accuracy, error value, memory, and time. The experimentation is performed by KDD dataset. Conclusion: A result shows that our proposed system has maximum accuracy and attains the effective convolution process to improve privacy preserving.

Background: Cloud computing environment is a novel paradigm in which the services are hosted, delivered and managed over the internet. Tasks scheduling problem in the cloud has become a very interesting research area. However, the problem is more complex and challenging due to the dynamic nature of cloud and users’ needs as well as cloud providers’ requirements including the quality of service, users’ priorities and computing capabilities. Objective: The main objective is to solve the problem of tasks scheduling through an algorithm which can not only improves the client satisfaction, but also allows cloud service provider to gain maximum profit and ensure that the cloud resources are utilized efficiently. Method: (a) Optimization of the waiting time and the queue length. (b) Distribution of all requests into a novel queueing system in a dynamic manner based on a decision threshold. (c) Assignment of requests to VMs based on Particle Swarm Optimization and Simulated Annealing algorithms. (d) Incorporation of the priority constraint in the scheduling process by considering three priorities levels including the tasks, queues and VMs. Results: The results comparison of our algorithm with particle swarm optimization and First Come First Serve algorithms demonstrate the effectiveness of our algorithm in terms of waiting time, makespan, resources utilization and degree of imbalance. Conclusion: This study introduces an efficient strategy to schedule users’ tasks by using dynamic dispatch queues and particle swarm optimization with simulated annealing algorithms. Moreover, it incorporates the priority issue in the scheduling process.