Search

Facial Emotion Recognition (FER) is the approach of detecting emotions of humans from facial expressions. Emotions are detected automatically by the human brain and hence, many computer aided techn...

This paper mainly gives information about a compact 4-element multiple input multiple output (MIMO) antenna which is useful for IOT applications. Advantages of the proposed antenna is that it can be easily extendable for a large size array. The array consists of 4 elements placed at an angle of 90 degree to the adjacent element. The substrate used in the design is a low loss laminate of FR4 with dielectric constant of 4.4. Based on the results obtained the antenna covers a UWB band of 2.4 GHz–13.0 GHz and the isolations also exceed 20 dB. Return loss, radiation pattern and isolation is plotted. The simulation results show that the proposed antenna can perform well for MIMO devices. The proposed antenna is much suitable for the applications of IOT and RADAR.

Decision making by applying data mining methods is being used in many service organizations. Educational bodies gradually started to use the business intelligence techniques to identify the current progress in their institutions. Numerous factors which have an impact in academia will be vivid to the educationalists while applying data mining techniques on the academic data. By employing the data mining methodologies, we could identify different patterns which aid institutions to take strategic decisions to improve the students’ academic performance. Potential graduate students will have a dilemma on identifying the universities for their post graduate admissions and on the other hand an average graduate student would be uncertain on getting post graduate admission in a reputed university based on their academic scores. In this study, we applied the classification techniques such as Logistic Regression, KNN Classification, Support Vector Classification, Naive Bayes Classification, Decision Tree Classification and Random Forest Classification on the given academic admission dataset. By comparing the accuracy and mean absolute error of each model, the Logistic Regression classifier outperformed others with an accuracy of 99%.

Recently, the popularity for massive online open course (MOOC) learning among student community is increasing day by day. In most of the MOOC learning Web sites, indexing and retrieval techniques are based on the keywords like names, titles and Web addresses of the videos. As a result, the response for content-based queries is not good in most of the scenarios. In this paper, we propose an efficient scene content-based indexing and retrieval (ESCIR) framework for lecture videos. The proposed ESCIR framework consists of two phases: offline phase and online phase. In offline phase, we apply novel block-level key frame extraction (BLKFE) technique to segment video frames into shots and choose the right frame. The optical character recognition (OCR) tool is applied to the generated key frames in the videos for extracting the text. The jaccard similarity coefficient measure is used to eliminate the duplicate text frames, and then use of stop word removal and stemming algorithms are applied to get meaningful keywords from the scene. We used single-pass in-memory indexing (SPIMI) technique for building the indexing with the help of extracted keywords. In online phase, search and matching algorithms will map input queries given by user to the corresponding videos. We applied the okapi BestMatch25 (BM25) ranking function to rank those matched videos for best relevance results. The proposed ESCIR framework has been validated on standard video lectures and found to give better results than the existing state of art algorithms in terms of precision and accuracy. We compared our results with existing methods of lecture video techniques reflects improvising robust performance on lecture videos.

The image retrieval system is a computer system for browsing, searching and retrieving images from a large database of digital images or text. Most traditional and common methods of image retrieval utilize some method of adding metadata such as captioning, keywords or descriptions to the images so that retrieval can be performed over the annotation words. Manual image annotation is time-consuming, laborious and expensive. To address this, there has been a large amount of research done on automatic image annotation. With the rapid development of information technology, the number of electronic documents and digital content within documents exceeds the capacity of manual control and management. The usage of images is increased in real time. So, the proposed system concentrates on retrieving image by using the text-based image retrieval system. Text documents are given as input to the preprocessing stage, and features are extracted using TF-IDF. Finally, document clustering method can be used to automatically group the retrieved documents into list of meaningful categories. Document clustering clusters the document of different domains and latent Dirichlet allocation (LDA), each document may be viewed as a mixture of various topics where each document is considered to have a set of topics, and after that relevant documents are retrieved and then images of those relevant documents are retrieved.

One of the main challenges in the Traffic Anomaly Detection (TAD) system is the ability to deal with unknown target scenes. As a result, the TAD system performs less in detecting anomalies. This paper introduces a novelty in the form of Adaptive Neuro-Fuzzy Inference System-Lossy-Count-based Topic Extraction (ANFIS-LCTE) for classification of anomalies in source and target traffic scenes. The process of transforming the input variables, learning the semantic rules in source scene and transferring the model to target scene achieves the transfer learning property. The proposed ANFIS-LCTE transfer learning model consists of four steps. (1) Low level visual items are extracted only for motion regions using optical flow technique. (2) Temporal transactions are created using aggregation of visual items for each set of frames. (3) An LCTE is applied for each set of temporal transaction to extract latent sequential topics. (4) ANFIS training is done with the back-propagation gradient descent method. The proposed ANFIS model framework is tested on standard dataset and performance is evaluated in terms of training performance and classification accuracies. Experimental results confirm that the proposed ANFIS-LCTE approach performs well in both source and target datasets.

Federated clouds have been a solution to some of the challenges of cloud computing like vendor lock-in and performance-related issues in terms of a wide range of resource utilization and pricing for cloud consumers. This paper provides much insight into the problems faced by cloud consumers while utilizing resources for particular price in relation to SLA violation, QoS awareness and cloud brokerage. A brief review of resource utilization with pricing in perspective of cloud consumers is presented, and a layered agent-based model was proposed for simulating federated cloud. To analyze maximum resource utilization on pricing option a MaxResourceUtility, an expected maximization (EM) algorithm was proposed to consider the influence of missing QoS factors while estimating it for resource utility. The results show that 5–10% increase in maximum resource utility and 10–20% decrease in pricing are observed by considering QoS factor response time while utilizing resources.

The amount of digital data has been increasing exponentially. There is a need to reduce the amount of storage space by storing data efficiently. De-duplication tends to save a lot of storage space in commercial applications like backup. De-duplication suffers from a serious disadvantage that data cannot be referenced across various locations. Soft links can be established only within a single disk that has been used for storage. In this paper we discuss cross level links that can be established across data which in turn is stored in a distributive manner. When user level data is stored across various volumes separated geographically a medium is required to reference data at the same time saving storage space. We implemented at file level de-duplication, a significant reduction in storage space was observed.

In this paper we present an approach to automatically detect anomalous traffic events like pedestrians crossing the junction, based on traffic video of low-level features such as size of the blob, spatial location, and velocity. The construction of Petri-Nets was used for both semantic description and event detection within traffic videos. The major novelties of this paper are extensions to both the modeling and the recognition capacities of Object Petri-Nets (PN). The detection of object level features are done with the help of state of art techniques like Gaussian Mixture of Models (GMM), and a series of Petri-Nets composed of various objects is proposed to describe the video content. The expected outcome of the proposed framework is that we can easily build semantic detectors based on PNs to search within traffic videos and identify interesting events. Experimental results based on recorded traffic video data sets and synthetic data sets are used to illustrate the potential of this framework.

As more and more surveillance cameras are deployed in a facility or area the demand for automatic detection of suspicious objects is increasing. Most of the work in recent literature concentrated on protruding object detection in video sequences. This paper proposes a novel approach to detect protruding as well as non protruding objects in sequences of walking pedestrians based on texture of the foreground objects. Initially static background is modeled with the help of mixture of Gaussian algorithm and the foreground objects are segmented. Later object is detected frame by frame which is followed by the calculation of statistical parameters such as mean and standard deviation, in every blob, to form data streams. These parameters are clustered online using k-means methodology over data streams, in order to find the outliers (dropped objects). Here k is based on the number of objects present in the video. Finally we have implemented on a standard data set from the website Video Surveillance Online Repository [15] and also our own dataset. The experimental results show that our system performs reasonable well and can accurately detect dropped objects in video data streams.

As powerful computers and cameras have become wide spread, the number of applications using vision techniques has increased enormously. One such application that has received significant attention from the computer vision community is traffic surveillance. We propose a new event detection technique for detecting abnormal events in traffic video surveillance. The main objective of this work is to detect the abnormal events which normally occur at junction, in video surveillance. Our work consists of two phases 1) Training Phase 2) Testing Phase. Our main novelty in this work is modified lossy counting algorithm based on set approach. Initially, the frames are divided into grid regions at the junction and labels are assigned. The proposed work consist of blob detection and tracking, conversion of object location to data streams, frequent item set mining and pattern matching. In the training phase, blob detection is carried out by separating the modelled static background frame using Gaussian mixture models (GMM) and this will be carried out for every frame for tracking purpose. The blobs location is determined by assigning to the corresponding grid label and numbered moving object direction to form data streams. A modified lossy counting algorithm is performed over temporal data steams for discovering regular spatial video patterns. In testing phase, the same process is repeated except frequent item set mining, for finding the spatial pattern in each frame and it is compared with stored regular video patterns for abnormal event detection. The proposed system has shown significant improvement in performance over to the existing techniques.

A 5 kVA, 415 V, 48–52 Hz, three-phase laboratory model Flexible AC Transmission System (FACTS) built around intelligent microprocessor based thyristor controlled switched capacitor is described. This laboratory model has been developed to conduct the feasibility studies for implementation of rapid impedance control on a 400 kV double circuit transmission corridor between Itarsi and Indore in the central part of India. The laboratory model uses scaled parameters of this transmission corridor for maintaining the receiving end voltage at Indore. The system studies reported in this paper include the simulation on PC and also the intelligent control aspect eg fault diagnosis and its identification, along with other steady state experimental investigations made on the newly developed laboratory model controller. A comparison between the simulated and experimental results reveals good correlation.

Career assessment is useful for a student in order to know the suitable jobs for him/her in future basing on student’s knowledge memory power, Intelligence, psychological aptitude, etc. During the academic program of a student, it is very important to assess the appropriate career(s), so that a student can select some appropriate electives and some specialized subjects, which lead to an appropriate career(s). In the process of academic program of a student, it is highly essential to plan his/her career. Generally, the career counselor in an institution analyzes the student’s academic record/Cumulative Grade Point Average (C.G.P.A.) and predicts suitable career(s) in industry. In case of students belonging to Computer Science and Engineering branch, counselor will suggest some appropriate jobs in software industry viz. software designer, software engineer, tester, marketing person, etc. based on the academic record. Apart from academic record, one should also consider the student’s psychological factors like intelligence, problem solving ability, patience, etc. to predict a better career. We developed an expert system for predicting career by assessing psychological factors of a student like problem solving ability, intelligence and patience levels of a student. In order to assess these psychological factors, we developed 8-puzzle game to assess student’s intelligence and planning ability levels, fastness in playing game, and patience levels. The system requests the student to play 8-puzzle game many times and displays the scores of student. A career table is designed which consists of list of careers in software industry and the minimum levels of requirements in academic record, intelligence, patience, etc. The academic record and psychological factors of the student are compared with the minimum levels required for each career and thus the system predicts the matching career(s) for the student.