Search

A common problem in pixelwise classification or semantic segmentation is class imbalance, which tends to reduce the classification accuracy of minority-class regions. An effective way to address this is to tune the loss function, particularly when Cross Entropy (CE), is used for classification. Although several CE variants have been reported in previous studies to address this problem, for example, Weighted Cross Entropy (WCE), Dual Cross Entropy (DCE), and Focal Loss (FL), each has their own limitations, such as introducing a vanishing gradient, penalizing negative classes inversely, or a sub-optimal loss weighting between classes. This limits their ability to improve classification accuracy or reduces their ease of use. Focal Loss has proven to be effective at balancing loss by increasing the loss on hard-to-classify classes. However, it tends to produce a vanishing gradient during backpropagation. To address these limitations, a Dual Focal Loss (DFL) function is proposed to improve the classification accuracy of the unbalanced classes in a dataset. The proposed loss function modifies the loss scaling method of FL to be effective against a vanishing gradient. In addition, inspired by DCE, a regularization term has also been added to DFL to constrain the negative class labels to further reduce the vanishing gradient effect and increase the loss on hard-to-classify classes. Experimental results show that DFL has better training performance, and provides greater accuracy compared to CE, WCE, FL and DCE in every test run conducted over a variety of different network models and datasets.

Stroke is a drawn-out incapacity illness caused everywhere on the world and it is the third driving reason for demise. Pre-syndromes of stroke dives more importance now a day. Stroke happened basically because of individual way of life in the cutting the edge period changing components, for example, high glucose, coronary illness, heftiness, diabetes. In this research, we analyse different Supervised machine Learning algorithms (i.e. SVC, DT, RFC, Logistic Regression, Linear Regression, Bernoulli NB, Gaussian NB and K-Nearest Neighbours Classifier). The accuracy in the work have found highest by applying Decision Tree classifier which is 0. 93.

Since wind power is directly influenced by wind speed, long-term wind speed forecasting (WSF) plays an important role for wind farm installation. WSF is essential for controlling, energy management and scheduled wind power generation in wind farm. The proposed investigation in this paper provides 30-days-ahead WSF. Nonlinear Autoregressive (NAR) and Nonlinear Autoregressive Exogenous (NARX) Neural Network (NN) with different network settings have been used to facilitate the wind power generation. The essence of this study is that it compares the effect of activation functions (namely, tansig and logsig) in the performance of time series forecasting since activation function is the core element of any artificial neural network model. A set of wind speed data was collected from different meteorological stations in Malaysia, situated in Kuala Lumpur, Kuantan, and Melaka. The proposed activation functions tansig of NARNN and NARXNN resulted in promising outcomes in terms of very small error between actual and predicted wind speed as well as the comparison for the logsig transfer function results.

The rapid increase of wireless technologies and body area networks has enabled the continuous healthcare monitoring of patients from a remote location using various sensors. This paper describes a remote health monitoring system using WBAN where different sensors are used to collect a patient’s vital signs. Once the data is received, the captured data are evaluated by the Arduino and sent to the patient’s Android Smartphone using a Bluetooth module and a newly developed android application named Health Monitoring App. Data can be accessed by the patient himself, by doctor or patient’s family members. In critical conditions such as when body temperature and pulse rate goes beyond a predefined value or patient fall down, then application from patient’s smart phone triggers a message to deliver to the doctor or relatives with the location of the patient using built-in GPS in patient’s phone. This remote health monitoring system is very feasible and cost-effective for a developing country like Bangladesh.

Researchers around the world are now focused on to make our devices more interactive and trying to make the devices operational with minimal physical contact. In this research, we propose an interactive computer system which can operate without any physical keyboard and mouse. This system can be beneficial to everyone, especially to the paralyzed people who face difficulties to operate physical keyboard and mouse. We used computer vision so that user can type on virtual keyboard using a yellow-colored cap on his fingertip, and can also navigate to mouse controlling system. Once the user is in mouse controlling mode, user can perform all the mouse operations only by showing different number of fingers. We validated both module of our system by a 52 years old paralyzed person and achieved around 80% accuracy on average. ISBN för värdpublikation: 978-3-030-68153-1, 978-3-030-68154-8

Continuous launching of new satellites and increasing numbers of space missions is making space a congested environment. Collision with space debris or other satellites is now a real problem for satellites with the problem more acute in highly trafficked orbits. Thus, mission operators and space agencies are in need of high accuracy collision avoidance systems for spacecrafts and satellites. This paper focuses on tackling the satellite collision problem by implementing a collision avoidance system using neural networks and relevant machine learning techniques. The primary model is based on General Regression Neural Network (GRNN) and the secondary models are based on Artificial Neural Networks (ANN), Random Forest Regression & Support Vector Regression techniques. The dataset used in this paper is collected from the ESA (European Space Agency) which contains the events of risk assessment or in other words, Conjunction Data Messages (CDM). The proposed collision avoidance system predicts the collision risk percentage between target (a satellite of interest) and chaser (space debris or another satellite) objects. The predicted risk enables the target to maneuver accordingly and ultimately avoid collision with the chaser object. The GRNN algorithm uses lazy learning which does not require iterative training and makes predictions based on stored parameters. The training data has been normalized before applying the algorithm as GRNN network is sensitive to high deviation among input features. However, the GRNN model predicts the risk of collision between the target & the chaser object with an MSE (Mean Squared Error) of 11% which means the model predicts the risk of collision with 89% accuracy and this 89% risk can give enough confidence factor to the concerned authority to take necessary evasive maneuvers. This is reliable enough and lower than other models' MSE to consolidate the fact that the GRNN model is best fit for our dataset.

Nowadays, the recognition of facial expression draws significant attention in various domains. In view of this, a realtime facial expression recognition system has been developed using a Deep Learning approach, which can classify ten emotions, including angry, disgust, fear, happy, mockery, neutral, sad, surprise, think, and wink. In addition, an integrated expert system has also been developed by integrating Deep Learning with a Belief Rule Base to support the assessment of the overall mental state of a person over a period of time from video streaming data under uncertainty. In this research, data-driven and knowledge-driven approaches are integrated together to assess the mental state of an individual. Such a system could enable the identification of a suspect before committing any crime beforehand by the law enforcement agency. The performance of this integrated system is found reliable than existing methods of facial expression assessment. Contribution- The paper presents a noble method of computing the overall mental condition of a person by integrating CNN and BRBES under uncertainty. Contribution- The paper presents a noble method of computing the overall mental condition of a person by integrating CNN and BRBES under uncertainty. ISBN för värdpublikation: 978-1-7281-9331-1

In this study, Random Forest Regressor, Linear Regression, Generalized Regression Neural Network (GRNN) and Fully connected Neural Network (FCNN) models are leveraged for predicting unconfined compression coefficient with respect to standard penetration test (N-value), depth and soil type. The study is focused on a particular correlation of undrained shear strength of clay (Cu) with the standard penetration strength. The data used is from 14 no. ward in Mymensingh and Rangamati districts which are situated in Bangladesh. By using this data, the study tries to solidify the correlation of SPT (N-value) with Cu. It also tries to check the goodness of the relationship by comparing it with unconfined compression strength values gained from the unconfined compression test calculated from the field by experts.

Application of Radial Basis Function Network (RBFN) to impact force identification.Application of Multilayer Perceptron (MLP) to the same task.Accuracy, success rate and error range comparison between RBFN and MLP.Estimation of accuracy improvement for using RBFN in place of MLP. Impact force identification from response sensors is important especially when force measurement using force sensor is not possible due to the installation or dynamic characteristic altering problems. For example, the bump-excited impact force acting on vehicle wheel or ship collision on an offshore structure. Among various existing impact identification approaches, neural network based force identification method has received great attention because one does not need to have a system model. Thus, it is less likely to be affected by ill-posed problem that often occurs during the inversion process. So far, previous studies focused on solving the impact force identification problem using only the conventional Multilayer Perceptron (MLP). Thus, there is a room for improvement to find an alternate algorithm that has great advantage over MLP. For this reason, this study proposes Radial Basis Function Network (RBFN) for possible further improvement in impact identification task. A comparative study between these two algorithms was conducted via experimental approach. Impact forces were made on a Perspex plate structure which was designed to produce similar dynamic behavior of a typical vehicle. Impact locations were fixed at four edges of the test rig to simulate impact events at a vehicle's wheels. Time-domain peak-to-peak and peak arrival time features were extracted from accelerometer data to use as network inputs. Few training data were taken in the way that they represent the entire range of magnitudes of all trial impacts made throughout the experiment. In overall, RBFN improved the impact localization and quantification accuracies by decreasing 32.98% and 40.91% error respectively compared to MLP. The improvement was mainly due to the RBFN's strong approximation ability and its superior tolerance to experimental noises/uncertainties.

In this world of ever-increasing demand for electrical power generation, the urge for a promising, sustainable and yet environment-friendly energy source has led us to pay more attention to utilizing the immense potential of ocean wave energy. Numerous researches till now shows that the efficiency of wave energy extraction depends highly on the shape and design of the device and also its interaction with the energy resource. This paper proposes a particular design of a floating buoy point absorber wave energy converter considering its feasibility in the geographical condition of the coastal region in Bangladesh. After discussing the design algorithm and its functionality, a procedure has been given to evaluate the wave energy converter energy production quantitatively with the help of WEC-Sim library on MATLAB software. As the ocean waves are irregular and the wave height and period varies significantly from wave to wave, the generated power is highly variable which imposes a big challenge in harnessing ocean wave energy for electricity generation. In this study, a method has been proposed to generate steady sinusoidal voltage from the variable electrical power produced from irregular ocean waves. Next, the available data on wave characteristics of several coastal regions have been examined and checked against the proposed design. Finally, analyzing the simulation results and considering the possibilities and limitations of energy production, some recommendations have been given on where such devices can be installed.

A precise credit risk assessment system is always vital to any financial institution for impeccable and gainful functioning. In such an ever-changing economy as the rate of loan defaults are gradually increasing, authorities of financial institutions are finding it more and more difficult to correctly assess loan requests and tackle the risks of loan defaulters. In light of these events this paper proposes a machine learning model which can precisely assess credit risk and predict possible loan defaulters for credit lending institutions. A comparative analysis has been made using tuned supervised learning algorithms such as Support Vector Machine, Random Forest, Extreme Gradient Boosting and Logistic Regression for identifying defaulters. Recursive Feature Elimination with Cross-Validation and Principal Component Analysis have been used for dimensionality reduction. Metrics such as F1 score, AUC score, prediction accuracy, precision and recall have been used to evaluate each model. Among all the models, the combination of a tuned Support Vector Machine and Recursive Feature Elimination with Cross-Validation have shown great promise in identifying loan defaulters. The proposed model, therefore, can assist financial institutions in accurately identifying loan defaulters and prevent them from incurring further loss.

In the society, most of the disabled people especially completely movement impaired ones live like burdens and suffer from severe depression as they have to depend on others for day-to-day activities and cannot contribute to the community. These people may not be able to move their arms or legs but they can certainly think. If their ability to think and make decisions can be utilized in decision making situations and also if they can further be included in industrial automation structure, they will no longer be burden rather they will become man-power to the society. One of the major challenges of involving these people in such processes is making a low cost device that is affordable by the mass, made with locally available components and also interprets the person's choice accurately. As these people are physically disable, comprehending their electroencephalogram (EEG) signal for blink detection can be an effective solution for translating their choices. EEG-based controlling devices are mobile and can serve as powerful aids for severely disabled people in their daily life. In this work, a low cost EEG sensor is made with readily available components and a framework is developed to detect voluntary eye blinks. A graphical user interface (GUI) is also developed which is suitable for the impaired person. The eye blinks of the person act as bridge between a set of choices pointed in the GUI and his/her selection interest and thus help the person control home appliances, wheel chair and even computer without depending on others.

Millimeter-Wave Transformer Coupled amplifier circuit belongs input-output matching network, Impedance transforming network (ITN) with other connecting passive elements. Both matching and ITN need to be perfectly optimized to design schematic and layout finally. Generally, ideal and practical circuit extracted parameters are remained same for the best matching of amplifiers, mixer and Voltage Controlled Oscillator (VCO). However, Transformer Coupled matching networks are quite different regarding to practical circuit parameters optimization. Ideally, Transformer circuit optimization is very difficult to extract parameters due to generating leakage and parasitic capacitance and resistance. So, mm-Wave transformers are being transformed into Tee-Type and Pi-Type circuit shape and optimized practically based on their physical phenomenon and goal functions. Higher quality factor and self-resonance frequency, lower insertion loss, parasitic inter-winding capacitance minimization and lower internal parasitic capacitance and resistance of each different transformers shape (Octagonal Planar and Stacked) have been examined for mm-wave amplifiers design. This amplifier module may be included in Radar TRX and applicable for automotive vehicle Communication.

This paper presents a way to track eye-gaze by using webcam and mapping the eye-gaze data compensating head pose and orientation on the display screen. First, we have shown a blank screen with red dots to 10 individuals and recorded their eye-gaze pattern and head orientation associated with that screen location by automated annotation. Then, we trained a neural network to learn the relationship between eye-gaze and head pose with screen location. The proposed method can map eye-gazes to screen with 68.3% accuracy. Next, by using the trained model to estimate eye gaze on screen, we have evaluated content of a website. This gives us an automated way to evaluate the UI of a website. The evaluation metric might be used with several other metrics to define a standard for web design and layout. This also gives insight to the likes and dislikes, important areas of a website. Also, eye tracking by only a webcam simplifies the matter to use this technology in various fields which might open the future prospect of enormous applications.

Detecting emotion from facial expression has become an urgent need because of its immense applications in artificial intelligence such as human-computer collaboration, data-driven animation, human-robot communication etc. Since it is a demanding and interesting problem in computer vision, several works had been conducted regarding this topic. The objective of this research is to develop a facial expression recognition system based on convolutional neural network with data augmentation. This approach enables to classify seven basic emotions consist of angry, disgust, fear, happy, neutral, sad and surprise from image data. Convolutional neural network with data augmentation leads to higher validation accuracy than the other existing models (which is 96.24%) as well as helps to overcome their limitations. ISBN för värdpublikation: 978-1-7281-0788-2, 978-1-7281-0786-8 A belief-rule-based DSS to assess flood risks by using wireless sensor networks

In this paper, the design and implementation of SENSEP ACK - a smart system to monitor room environment of mushroom cultivation farm is reported. The system measures the temperature, humidity, light and CO 2 level with appropriate sensors. The obtained data can be utilized to control the environment of the nursing room with the help of external devices such as water pump, exhaust fan, light bulb and air conditioner. An android/web-based application has been developed for the ease of monitoring the system. The software application allows automation and remote operation of the external devices to maintain and control the optimum environment.

This paper proposes a process of sentiment analysis of movie reviews written in Bangla language. This process can automate the analysis of audience’s reaction towards a specific movie or TV show. With more and more people expressing their opinions openly in the social networking sites, analyzing the sentiment of comments made about a specific movie can indicate how well the movie is being accepted by the general public. The dataset used in this experiment was collected and labeled manually from publicly available comments and posts from social media websites. Using Support Vector Machine algorithm, this model achieves 88.90% accuracy on the test set and by using Long Short Term Memory network [1] the model manages to achieve 82.42% accuracy. Furthermore, a comparison with some other machine learning approaches is presented in this paper. ISBN för värdpublikation: 978-1-7281-5241-7, 978-1-7281-5242-4

This paper proposes a process of Handwritten Character Recognition to recognize and convert images of individual Bangla handwritten characters into electronically editable format, which will create opportunities for further research and can also have various practical applications. The dataset used in this experiment is the BanglaLekha-Isolated dataset [1]. Using Convolutional Neural Network, this model achieves 91.81% accuracy on the alphabets (50 character classes) on the base dataset, and after expanding the number of images to 200,000 using data augmentation, the accuracy achieved on the test set is 95.25%. The model was hosted on a web server for the ease of testing and interaction with the model. Furthermore, a comparison with other machine learning approaches is presented. ISBN för värdpublikation: 978-1-7281-0788-2, 978-1-7281-0786-8 A belief-rule-based DSS to assess flood risks by using wireless sensor networks

Though speech recognition has been a common interest of researchers over the last couple of decades, but very few works have been done on Bangla voice recognition. In this research, we developed a digital personal assistant for handicapped people which recognizes continuous Bangla voice commands. We employed the cross-correlation technique which compares the energy of Bangla voice commands with prerecorded reference signals. After recognizing a Bangla command, it executes a task specified by that command. Mouse cursor can also be controlled using the facial movement of a user. We validated our model in three different environments (noisy, moderate and noiseless) so that the model can act naturally. We also compared our proposed model with a combined model of MFCC & DTW, and another model which combines crosscorrelation with LPC. Results indicate that the proposed model achieves a huge accuracy and smaller response time comparing to the other two techniques. ISBN för värdpublikation: 978-1-7281-5852-5, 978-1-7281-5853-2

Patient monitoring is a pivotal part of the health care system nowadays, either at hospitals or at home. This paper proposes an intelligent patient monitoring system that automatically screens the patient’s health condition through various sensors. The data is then processed using a Raspberry Pi and useful information is saved to the IoT cloud. Primarily the system would be extracting the bio signal, ECG using an ECG sensor. Through continuous monitoring and graphical representation of the patient’s information, doctors/nurses/relatives can remotely check the patient’s condition. Furthermore, if the condition becomes critical, a notification is sent to the doctor/nurse/relative to inform them and either party will have the opportunity to start a video call.

Cloud federation is introduced to eliminate the resource limitation problem of individual Cloud Service Provider (CSP). In a federation, one CSP can outsource their overhead requests by hiring unused resources from other CSPs in the federation and CSPs cash greater revenue. In this paper, we propose a system that maximizes this revenue while putting some factor constraints in satisfactory standard. Factors include response time of instances that is analogous to Quality of Service (QoS), and profit that maintains a certain reputation. Response time is reduced by considering geographical location of the user and the data center of desired resources. In this system, CSPs choose suitable resources from resource pool of cloud federation that maximizes the profit while preserving Quality of Service (QoS) high and minimizing the rejection rate of user requests. We map factor constraints to MMKP (Multi-choice Multidimensional Knapsack Problem) algorithm and define objective function for revenue maximization. We calculate response time by Cloud Analyst to measure the QoS. Experimental results show that the proposed system can effectively increases QoS level by reducing response time and rejection requests in a cloud federation that eventually increases the revenue of the CSP.

The occurrence rate of death and injury due to road traffic accidents is rising increasingly globally day by day. For several decades, the focus of research has been on getting a deeper understanding of the significant factors that influence the risk of road traffic fatalities. In today's modern world, neural network (NN) approaches play a crucial role in identifying the contributing factors that describe the frequency and severity of road accidents. Over the years, many researchers used neural network models for predicting the impact of such factors on road accident injury severity. Deep learning methods such as the recurrent neural network (RNN) and the convolutional neural network (CNN) has recently been successfully used for the prediction of road accidents and demonstrate their high accuracy and efficiency. This study overview and summarizes the different forms of neural network models such as the single layer perceptron (SLP) neural network, the multilayer layer perceptron (MLP) neural network, the radial basis function (RBF) neural network, the recurrent neural network, and the convolutional neural network used as a prediction method for the severity of road crash injuries and includes a discussion of future planning and difficulties. This article also summarizes the model input parameter or independent variable and output or dependent variable, as well as various performance assessment methods.

The diagnosis and treatment of prostate cancer requires the accurate segmentation of the prostate in Magnetic Resonance Images (MRI). Manual segmentation is currently the most accurate method of performing this task. However, this requires specialist knowledge, and is time consuming. To overcome these limitations, we demonstrate an automatic segmentation of the prostate region in MRI images using a VGG19-based fully convolutional neural network. This new network, VGG19RSeg, identifies a region of interest in the image using semantic segmentation, that is, a pixel-wise classification of the content of the input image. Although several studies have applied fully convolutional neural networks to medical image segmentation tasks, our study introduces two new forms of residual connections (remote and neighbouring) which increases the accuracy of segmentation over the basic architecture. Our results, using this new architecture, show that the proposed VGG19RSeg can achieve a mean Dice Similarity Coefficient of 94.57%, making it more accurate than comparable methods reported in the literature.

According to statistics, the number of active Android device user is approximately two billion and 25% of Android devices are dual SIM(Subscriber Identification Module) supportive throughout the world. For this huge number of dual SIM users, it is quite difficult to remember which SIM package offers the cheapest rate for calling. In this study, we propose an automatic cost reduction application which assists in reducing cost of outgoing call by selecting the best available SIM. This app selects the best cost-effective SIM and automatically transfers call to the particular SIM. If the selected SIM does not have enough money for calling, the call is automatically transferred to the other SIM. Our app is an approach towards saving the additional amount of money that is being spent by the dual SIM users. User evaluation of partial implementation of our system confirms around 45 percent saving of money while using this app.

Display Omitted ANNs-solved vibration based parametric identification studies are reviewed.Factors which affect identification result are discussed.Pros and cons of ANN approaches are mentioned.Suggestions are given to potential researchers based on the discussion.Analysis with experimental results is provided to justify some point of view. Vibration behavior of any solid structure reveals certain dynamic characteristics and property parameters of that structure. Inverse problems dealing with vibration response utilize the response signals to find out input factors and/or certain structural properties. Due to certain drawbacks of traditional solutions to inverse problems, ANNs have gained a major popularity in this field. This paper reviews some earlier researches where ANNs were applied to solve different vibration-based inverse parametric identification problems. The adoption of different ANN algorithms, input-output schemes and required signal processing were denoted in considerable detail. In addition, a number of issues have been reported, including the factors that affect ANNs prediction, as well as the advantage and disadvantage of ANN approaches with respect to general inverse methods Based on the critical analysis, suggestions to potential researchers have also been provided for future scopes.

We present a method of generation of one-time keys (OTK) for single line authentication using zero knowledge (ZK) computation as undertaken by an authentication client application on a mobile device and a registration server. The method comprises initiation of an activation process on the authentication client by a user associated with the mobile device, a ZK computation sequence on the authentication client for generation of storage key and transport key, transportation of the ZK computation outcome to the browser, user input into a text-entry element, and then additional server-side processing for association of the authentication client on the mobile device. This would be inclusive of ZK computations for the user identifier (UID), a user-associated hashed message authentication code (HMAC) key for subsequent authentication interactions, and additionally a transport key, for secure server-to-device transmission these user-specific parameters on an out-of-band (OOB) channel as presumed to be insecure. Thereafter, additional ZK computations are undertaken client-side for recovery of UID and the HMAC key, and following that insertion onto device-side secure storage such that correct recovery can only be undertaken on the particular mobile device on which computations were previously undertaken.

This paper proposes a new technique for DNA cryptography that uses dynamic DNA sequence table to enhance the level of security. While handling with secure data, the requirements like compression, speed-up computation and processing etc are crucial issues. Bio-molecular DNA features possess the capability to cope up with these requirements. Existing DNA cryptographic techniques usually consider fixed DNA sequence table i.e., DNA bases and thereby the security is suspected to be breached by the intruder. To overcome this limitation, the proposed technique considers dynamic sequence table that assigns random ASCII characters to DNA sequence table initially. Then a finite number of iterations are applied based on a mathematical series where in every iteration the positions of ASCII characters are changed dynamically in the sequence table. Later on, One-Time-Pad (OTP) is applied on the modified encoding binary value. Again OTP ciphertext is processed through genomic conversion. Finally, it is converted into compressed ciphertext using amino acid table consisting of protein sequence that increases the confusion of the ciphertext. At last, the time requirements for encoding-decoding and encryption-decryption are evaluated and comparisons with other DNA techniques are presented.

Data mining is the procedure of breaking down data from unlike perspectives and resuming it into useful information. It is very important in the field of classification of the objects. It has been fruitfully applied in expert systems to get knowledge. We can determine appropriate classification of unknown objects according to decision tree rules by applying inductive methods to the given values of attributes of those objects. In this paper a decision tree learning algorithm ID3 is applied to build a decision tree in achieving our goal to gender identification of unknown objects. Our experiments have used records of 50 individuals among which 26 were male and 24 were female subjects having age groups of 19 to 25 years. The classification related to the training sets is done by proper calculation. The output of the work is the classified decision tree and the decision rules. It has been observed that the proposed decision tree can recognize 45 subjects gender from 50 individuals. It is a faster process in recognition of individuals' gender and having accuracy level 85% to 90%.

A model of full duplex hybrid CO-OFDM-TWDM PON integrated with Wireless Local Area Network (WLAN) is proposed here which is an emerging optical-wireless (Fi-Wi) network to enhance the facility of long haul internet and data transmission system. In the proposed model, the hybrid PON gets ameliorated than the all Hybrid TDM-WDM PON, OFDM-WDM PON. Due to its internal structures the Optical Line Terminal get attached with a large amount of Optical Network Unit through perfect optical signal processing technique and high speed optical access devices. The WLAN become easily matched with hybrid PON because of its quick power conversion capability and just enhanced downstream/upstream signals toward all TX/RX prototypes. The proposed model will take a better explication to increase recent release 12 of LTE-A in Fi-Wi network development of optical-wireless communication system. This sort of Fi-Wi network may develop the LTE-Advanced radio over fiber network as well.

Finger prints are considered as the best tool of identification. Gender identification from fingerprints is an important step in forensic anthropology in order to identify the gender of a criminal and minimize the list of suspect search. Fingerprint identification and classification has been extensively researched in the literature however very few researchers have studied the fingerprint gender classification problem. In the present work, we proposed an easy and efficient technique to classify gender of a person using systematic fingerprint ridge counting to a valley area. The study was carried out on the age group of 18 to 40 years where the students, workers, housewives etc are taken as the subjects. An optimal threshold for each transform is chosen for better results. It is found that the proposed system produces accurate decision of 85% of male and 74% of female. The overall performance of the system is found to be satisfactory and useful to forensic anthropology.

Optical-Wireless technology recently has emphasized the cellular network over ground and underwater communication. Passive optical network (PON) is the faster data operating platform and gets inter-linked with the wireless function to enhance its performance. Various PON systems setup is perfect to reduce the channel noise, interference, chromatic (CD), and polarization mode dispersion (PMD). This optical signal gets converted to electrical domain and finally photonic MIMO Micro-strip antennas propagate wireless data to user accurately. Here 3D Massive MIMO antennas transmit beam through 3D channel with the help of PON system. In this model, the combinations of two system increase data transmission rates and identify the multi-user location simultaneously according to their pilot signal receiving. And this analysis may improve the recent 4G LTE-B network more powerful for multi-user communication system.

Full duplex Coherent Optical Orthogonal Frequency Division Multiplexing Multiple Input Multiple Output (CO-OFDM MIMO) is an emerging technology for high speed data transmission technique of 4 th generation Radio over Fiber (RoF) Network. In this paper we analyze a model of CO-OFDM data from optical transmitting laser which emit terahertz range signal through fiber reducing chromatic and Polarization mode dispersion. At receiving part, Optical signal do focus on graphene layered photo-detector and get immediate conversion to electrical pulse. In Multi-user CO-OFDM data communication via MIMO array transceiver through Multimode Fiber (MMF) which utilizes Dense Wavelength Division Multiplexing (DWDM) technique to make the Optical communication links more efficient and speedy. This analysis is for building block to design a graphene nano-photonics based transceiver module that take an important contribution in Long Term Evolution Advanced (LTE-A) Optical-Wireless signal Communication.

Steganography is the art or practice of concealing secrete information into a cover media like text, protocol, file etc. In our proposed algorithm, a new Steganographic technique is developed to hide large amount of data in Bitmap image using the concept of LSB verification method. This method uses first three color components to detect the actual position of the secret bit into the fourth color for embedding and extracting data. This method uses adjacent three colors of image pixels to define the embedding position of message bit into the forth color.

The identification of an individual finds diverse applications in many fields such as justice, crime scene, voter identification as well as vehicle driver license identification and so on. Several classic methods have already developed for identifying a person, which are complex, expensive and not flexible in many circumstances. Due to emergence and wide spread availability of portable smart phones it is the demand of time to develop such an application in mobile platform by overcoming the drawbacks of classical methods. This paper solves this problem using available mobile phones in the market.

Social network sites (SNS) have made strong position in the heart of millions of users as if it is their own society. Numerous SNS are available on the web and most of them are strongly motivated to interconnect a group of people. These sites suffer too much about theft of personal detail, privacy issues and disclosure of daily activities towards unexpected person. This might be happened for two reasons; first users may disclose personal detail or sites do not have adequate measures to protect this affair. In social network, the way we interact with each other should be just like real social life; intimacy and visualization should differ from level to level. We have developed one abstract solution where each person can expose and exchange his expression with other subject to individual role; we call it role based social network (RBSN) which is unique for using ontology to improve its semantic nature. We have also analyzed some flaws of conventional social networks with their features (e.g. bizarre relation assignment, generic suggestions, and limited privilege to expose heartfelt expression).

There are several researches have been conducted and different methods have been applied to extend the features of indoor location tracking using wireless devices. Fingerprint algorithm is one of the popular methods that have been used to track mobile nodes in indoor structure. This paper focuses on the enrichment of positioning accuracy of the Fingerprint algorithm. RSSI mechanism has been used on the Fingerprint Algorithm to encompass better performance for the accurate positioning. The paper has been proposed two different types of Fingerprint methods and their performance variants on different circumstances of indoor environment. The proposed Fingerprint Database mapping has been divided according to the RSSI data redundancy. The proposed Fingerprint methods has been successfully implemented and experimented on UTMMIMOS CoE lab. The experimented results of two Fingerprint methods has been analyzed and compared as well.

In this paper, we have proposed a new design technique of BCD Adder using newly constructed reversible gates are based on NMOS with pass transistor gates, where the conventional reversible gates are based on CMOS with transmission gates. We also compare the proposed reversible gates with the conventional CMOS reversible gates which show that the required number of Transistors is significantly reduced. &nbsp

The structure of a software system deteriorates as a result of continuous maintenance activity. For the purpose of software reengineering or reverse engineering, software engineers often get the original source code as the most updated source of information due to lack of current documentation and limited or nonexistent availability of the original designers. The application of clustering techniques to software systems aiming to discover feature-oriented and meaningful subsystems can help software engineers involved in software reengineering or reverse engineering to understand high-level features provided by those subsystems. Continuous research is going on in the recent years — addressing different issues in the software clustering problem. Our software clustering approach introduces the use of Knowledgebase, which leads to considerable improvement than the existing approaches. Similarity measurement is the key to perform successful clustering. Similarity measurement in the existing approaches has a common drawback that they do not incorporate the diversity of software systems. Our approach uses Knowledgebase which acts as a repository of information about the internal structure of the generic types of the software systems to provide guidelines on similarity measurement criteria and their respective weightages. The final clustering is done by populating automatically generated subsystems along with the known subsystems (provided by Knowledgebase). In our research, we have developed a tool named “ULAB Cluster 1.0” which implements our new clustering approach. This clustering tool has been evaluated by using a benchmark named “MoJo distance” for different well-known software systems. The experimental results show that our approach generates more appropriate subsystems than the other existing clustering approaches and outperforms others in different dimensions of software clustering quality.

Due to significant popularity of Location-based Services and Multimedia communication over mobile devices, there are many researches has been conducted to extend the features of location tracking and make it cost-effective to users. It becomes necessary for the users to have seamless communication with automated switching of software applications. This paper focuses on the performance of indoor location tracking system on IPv6 Network Island with multiple real time applications that has location assisted transfer for mobile users. RSSI (Received Signal Strength Indicator) mechanism has been used to locate the moving nodes. The developed location tracking server is having dynamic and centralized MySQL database management system. SIP (Session Initial Protocols) user agent has been used to deploy intercommunicating of multimedia data for instance; video and audio conference, text messaging among the moving nodes and user can transfer the seamlessly transfer the session to their nearest mobile node which will be determined by the Location Server. This paper is going to discuss about the seamless performance of SIP during the session transference. The developed project is cost-effective and precisely conducive for the industries or any indoor organization. The prototype of the project has been successfully developed and has been tested as well. The results show the seamless connectivity of the multimedia application during device switching.

Representation of indoor 3D location tracking had been an advantage to globally supported to track the mobile unit everywhere, even in different level of building. Our main focus is presenting development of indoor 3D location tracking to replace the existing 2D location tracking. The 2D location tracking mechanism only supports tracking the mobile unit position in the same level and provides handover or device switching. If the users need to go to another level, user will have to disconnect the connection. The development of the system that has a location tracking mechanism responsible to track the position of the mobile unit (such as laptop and PDA). In this paper, the system running in IPv4 framework and set up as a read test bed. The proposed location tracking will be based on Received Signal Strength Indicator (RSSI). The 3D location tracking system will be purely software based with minimum hardware dependent.

The IP Multimedia Subsystem (IMS) delivers multimedia services using IP integrated voice and packet services. IMS also provide users seamless connectivity. In this paper a location tracking system for IMS is developed track the location of a device. The existing testbed that supports Session Initiated Protocol (SIP) had been developed in the UTM MIMOS Coe, Universiti Teknologi Malaysia. However it is using 2D location tracking mechanism. The 2D location tracking mechanism only supports tracking the mobile unit position in the same level and provides handover or device switching. However if the user need to go to another level, user will have to disconnected the connection. This paper proposed a 3D indoor location tracking for mobile units allow roaming with SIP support. The location tracking will in a way help the user by providing seamless mobility at all services level. This paper present the initial results for the 3D location tracking tested at one level.