Ninu Preetha N S, Brammya G, Ramya R, Praveena S, Binu D, Rajakumar B R
The channels used to convey the human emotions consider actions, behaviors, poses, facial expressions, and speech. In fact, the facial expressions take prominent role in our daily life to communicate to other people. An immense research has been carried out to analyze the relationship between the facial emotions and these channels. The goal of this paper is to develop a system for Facial Emotion Recognition (FER) that can analyze the elemental facial expressions of human, such as normal, smile, sad, surprise, anger, fear and disgust. The recognition process of the proposed FER system is categorized into four processes, such as pre-processing, feature extraction, feature selection, and classification. After preprocessing, SIFT-based feature extraction method is used to extract the features from the facial point. Further, a metaheuristic algorithm called Grey Wolf Optimization (GWO) is used to select the optimal features. Subsequently, GWO-based Neural Network (NN) is used to classify the emotions from the selected features. To the next of the implementation, an effective performance analysis of the proposed as well as the conventional methods like Convolutional Neural Network (CNN), NN- Levenberg–Marquardt (NN-LM), NN- Gradient descent (NN-GD), NN- Evolutionary Algorithm (NN-EA), NNFirefly (NN-FF), and NN-Particle Swarm Optimization (NN-PSO), is provided by evaluating few performance measures and thereby, the effectiveness of the proposed strategy over the conventional methods is validated.
Distributed Parallel Medical data clustering based on the proposed BatDol-Sparse FCM-based Map Reduce Framework
Suki Antely A, Jegatheeswari P, Bibin Prasad M, Vinolin V,Vinusha S, Rajakumar B R and BinuD
Parallel clustering serves as a platform for handling the big data that updates asseveral millions of medical data with time. The literature displays a number of clusteringalgorithms using map-reduce framework, but they did not assure the effective clusters such thatknowledge extraction becomes tough. With the aim to render a better and effective dataclustering method to analyze the big medical data arriving from distributed systems, this paperuses a new clustering method. The proposed method named as BatDolphin-based Sparse FuzzyC-Means (BatDol-Sparse FCM) clustering algorithm is proposed that paves way for the optimalselection of the cluster centroids. The distributed medical big data is managed using the Map-Reduce framework that is inbuilt with the BatDolphin-based Sparse Fuzzy C-Means algorithmsuch that the local and global clustering is executed. Implementation of proposed BatDol-SparseFCM algorithm is done by taking six different medical datasets and evaluated based on metricssuch as clustering accuracy (CA) and dice coefficient (DC). From the simulation results it isevident that, the proposed parallel clustering scheme provided better results than the existingalgorithms with the values of 0.96 and 0.9667 for CA and DC respectively.
Deer Hunting Optimization Algorithm: A new nature inspired meta-heuristic paradigm
Brammya G, Praveena S, NinuPreetha N S, Ramya R,Rajakumar B R and Binu D
This paper proposes a novel meta-heuristic algorithm, named DHOA, inspired from the hunting procedure of humans toward deer. Even though the activities of the huntersdiffer, the way of attacking the buck/deer is based on the hunting strategy they develop. The hunting strategy depends on the movement of two hunters in their best positions, termed asleader and successor. Accordingly, each hunter updates his position until they reach the buck. The experimental results reveal that the proposed DHOA provides competitive results whencompared with the state-of- the-art optimization algorithms, such as GWO, WOA, FF, PSO, and so on. The experimentation is carried out with 39 benchmark functions and threeengineering applications. Moreover, a specific application is exploited by integrating NN in DHOA (DHOA-NN), to show the efficiency of the proposed algorithm in the classification.The proposed algorithm experimented in real-time engineering applications, and the performance comparison with the existing optimization algorithms proves the superiority ofthe DHOA algorithm..
Optimization using Lion Algorithm: A Biological Inspiration from Lion’sSocial Behavior
Rajakumar B R
During the past decade, solving complex optimization problems with bio-inspiredoptimization algorithms, especially evolutionary computation-based and swarm intelligence-based algorithms has received substantial attention among practitioners and researchers. Inthis paper, a novel optimization algorithm on the basis of the lion’s unique social behaviour isdeveloped. Here, unique lifestyle of the lion has been the fundamental motivation for thedevelopment of this optimization algorithm. The two most popular lion’s social behavioursare Territorial defense and Territorial takeover. Moreover, the algorithm is experimented onunimodal and multimodal benchmark minimization functions and compared with leadingswarm intelligence such as Artificial Bee colony (ABC), Bacterial Foraging OptimizationAlgorithm (BFO), Cuckoo Search (CS), FireFly (FF), Group Search Optimization (GSO), Moth-flame Optimization (MFO), Particle Swarm Optimization, Simulated Annealing, andDragon Fly algorithm and other algorithms Biogeography-based optimization (BBO),Differential Evolution (DE), Genetic Algorithm (GA), Gravitational Search Algorithm(GSA), Grey Wolf Optimization (GWO), Harmony Search Algorithm (HSA), SimulatedAnnealing (SA), Whale Optimization Algorithm (WOA), Crow Search Algorithm (CrS). Theobtained results show that lion algorithm is competent over majority of the evolutionaryalgorithms and equivalent to few other algorithms.
Lion Algorithm on Standard and Large Scale Engineering Problems
Rajakumar B R
Engineering optimization problems have become complex owing to their nonlinear objectivefunctions, where a number of linear and/or nonlinear constraints have to be satisfied. Hence,conventional optimization methods such as gradient-based algorithms are often unable tooffer satisfactory solutions. During the last two decades, evolutionary algorithms haveattained increasing consideration as optimization techniques for complex engineeringproblems. However, solving these problems remains as challenging yet. Hence, this proposedwork intends to exploit a new algorithm known as Lion Optimization Algorithm (LA), whichis based on lion’s unique social behaviour. The algorithm mimics the two most popular lion’ssocial behaviours, called as Territorial defense and territorial takeover. Here, lions follow aunique defined process to drive weak lions and to keep the species strong enough over others.These entire processes are also modelled here and hence named as lion algorithm. Moreover,the performance of this algorithm on engineering problems like pressure vessel design, Geartrain design, spring tension design, three-bar truss design, tension/compression spring andwelded beam are rigorously analyzed and compared both in standard as well as large scales.
C-GWO: Trust included Cluster Head SelectionModel in WSN using Crossover influenced Grey WolfOptimization
R. Ramya, T. Angelin Deepa, G. Brammya, N.S. NinuPreetha, S. Praveena,B.R. Raja Kumar, D. Binu
In wireless sensor networks (WSNs), energy efficiency is considered as the vitalaspect since the deployed sensor nodes are battery-operated devices. In order to fulfil theneed of energy efficient data transmission, clustering based approaches are implemented viadata aggregation so that the balancing of energy consumption among the sensor nodes innetwork may be more satisfactory. Clustering is the most important approach or strategy forprolonging the lifetime of network in WSNs. It engages in sensor nodes grouping to formclusters and in each cluster a sensor node should behave as head of the cluster, which istechnically termed as ‘Cluster Head’. The role of this cluster head is: It collects the data orinformation from respective nodes in cluster and forwards the aggregated data to Base Station(BS). However, the selection of cluster head is a serious problem in WSN and yet that is notup to the mark. This paper intends to propose a new cluster head selection approach thatincorporates many criteria including Energy, Delay, Distance and Trust. Moreover, the paperalso highly concentrates on secure data transmission, and this is adorably satisfied throughthe concern of Trust measure. The proposed Crossover influenced Grey Wolf Optimization(C-GWO) compares its performance over other conventional methods in terms of normalizedenergy and number of alive nodes.