Computer Engineering

Making recognition more reliable under uncontrolled lighting conditions is one of the most important challenges for practical face recognition systems This paper proposes a system which can handle illumination problem of face recognition systems by using “Retinex and color constancy” algorithm. The Retinex and color constancy approach has been plugged with Elastic Bunch Graph Matching (EBGM). The proposed system has been tested on FERET database having more than 1000 face images. The experimental results demonstrate that performance of the proposed system is superior to the known systems with the increase in accuracy.

Cloud computing is emerging technology; it has Stored more amount of data. When accessing the technology, it has to face many problems like load balancing, task scheduling. The main problem is physical host in cloud data center are so overloaded. While it happens, the data center has imbalanced. In existing implementation approaches load balancing concepts. It has much complexity. For this problem, we have introduced Load balancing based Bayes theorem and clustering with some constraints.

In recent years communication network has been undergoing major changes in network architecture and services. Network must be designed for flexibility and adaptability with respect to access topology and routing. The system should operate well under the variety of unexpected condition and node failures. Under such situation the shortest path table must be updated properly. Many traditional algorithms were used has several drawbacks .In order to resolve that drawbacks, a New VLSI based routing network model is proposed for routing which can handle component failures. The learning algorithm for routing is developed using cost function to be minimized to find the optimized path. It is simulated and synthesized using Verilogpro and Leonardo spectrum. Also it has been analyzed using c.

Mobile Agent is an alternate strategy to compose a dignified distributed system. Mobile agent is a program that can transform one host to another host at time in places of their choices (Jump and Go). The comparison between the Mobile Agent platforms Grasshopper and JAMES Framework as a suitable architecture for distributed system. A comparison is made and a categorization is made based on the performance of the James and Grasshopper mobile agents, with an emphasis on the James and Grasshopper framework as a viable architecture for distributed systems for multi-agent organizations. The purpose of this research is to compare and contrast mobile agent platforms.

Importance of ergonomics study in manual materials handling rises from the prospective risks human workplace accidents and injuries. Specifically lifting the materials from one place to other, diverse activities such as pushing, pulling, lowering, and holding, turning and carrying of weights. A risk to many humans are considered to be the prime cause of hip pain, joint impairments, stress and strains, sprains, dislocation of the lumbar spine disc, hip bone fracture, joint inflammation, tear of muscle tissue, contusion, and nerves problem due to often leading activity limitation and workplace accidents. This types of activities leads to increased worker reimbursement and loss of productive man hours. About one third of all jobs in engineering, production and business involve human in Manual Material Handling work. A finite elements model analyze the stresses in human hip joints using Image processing techniques, soft computing like MAT Lab and ANSYS. The more effort is to be taken for data collection and also during finite element modeling. A biomechanical model proposes the development and optimizing the lifting posture for minimum effort. This model is used to predict the lifting capabilities of each and every individuals. Future study can be extended for loading of the muscles strength in human.

We have dealt with the Backend design of LMS algorithm in Cadence tools. The LMS algorithm is used for the purposes of Polyharmonic Power Calibrator, Active phase cancellation of Hostile radars, and Noise cancellation. We have implemented the algorithm into VLSI technology using Verilog HDL, designed, simulated, and synthesized it using Xilinx Spartan3 3s400pq208 and the backend design is done using Cadence tools. The backend design for the LMS adaptive filter is done using RTL-GDSII So C encounter system and the So C design of the LMS algorithm had a power consumption of 32mW, timing slack of about 14ps and the design frequency is about 4.2Mhz also it uses an area of about 9753microns with a core size of about Ratio (H/W) 0.9204437. It is found that the backend implementation to be more efficient than the other methods of implementation.

The level of pollution has increased with times by lot of factors like the increase in population, increased vehicle use, industrialisation and urbanisation which results in harmful effects on human being by directly affecting health of population exposed to it. IOT Based system for monitoring and controlling Air Pollution in which monitor the Air Quality at any Industry over a web server using sensors when the air quality goes down across a certain level, means when there are sufficient amount of harmful gases are present in the air like CO2, smoke, alcohol, benzene and NH3. It will show the air quality in PPM on the display screen and as well as on website so that we can handle and monitor it easily. After gathered the information about air quality, this information is send to the Air Pollution Control Officer and the owner of Industry through mail and message in order to take strict action.

Plant diseases are widespread by a variety of pests, weeds and pathogens and can have devastating effects on agriculture if not treated in time. Farmers face a myriad of challenges due to adequate water supply, premature rain, storage facilities, and diseases of multiple crops. Plant diseases are a major threat to farmers with enormous production and economic loss. Identifying disease on a few hectares of farmland can be a daunting task, even in the presence of modern technology. Accurate and rapid disease prediction for early crop disease treatment has proven to be productive for healthy crops as well, minimizing personal financial loss. Many studies use modern deep learning approaches to improve the accuracy and performance of object detection and identification systems. Crop diseases are a constant challenge for smallholders, threatening income and food security. The proliferation of smartphones and recent revolutions in computer vision models has created a method of image classification in agriculture. Convolutional neural networks (CNNs) are considered the cutting edge of image recognition and offer the possibility of making timely and clear diagnoses. This article examines the performance of a pre-trained ResNet 34 model in the detection of plant diseases. The developed model can be used as a web application to detect seven plant diseases from healthy leaf tissue. A dataset containing images of 8,685 leaves. Captured in a controlled environment and established for model training and validation. Verification results show that the proposed method can achieve an accuracy of 97.2% and an F1 score of over 96.5%. It demonstrates the technical feasibility of CNN in plant disease classification and points the way to AI solutions for smallholders.

This paper presents a novel method for detecting human faces in an image with complex backgrounds. The approach is based on visual information of the face from the template image and is commenced with the estimation of the face area in the given image. As the genetic algorithm is a computationally expensive process, the searching space for possible face regions is limited to possible facial features such as eyes, nose, mouth, and eyebrows so that the required timing is greatly reduced. In addition, the lighting effect and orientation of the faces are considered and solved in this method. Experiments demonstrate that this face detector provides promising results for the images of individuals which contain quite a high degree of variability in expression, pose, and facial details. Hybrid algorithm is proposed to solve combinatorial optimization problem by using Ant Colony and Genetic programming algorithms. Evolutionary process of Ant Colony Optimization algorithm adapts genetic operations to enhance ant movement towards solution state. The algorithm converges to the optimal final solution, by accumulating the most effective sub-solutions.

The massive amounts of data collected today by software in fields varying from academia to business and many other fields, is increasingly becoming a huge problem due to storage technologies not advancing fast enough to provide the performance scalability needed. This is even truer for data which are highly organized and require analysis while being stored in databases and being accessed by various applications simultaneously. As database vendors struggle to gain more market share new technologies emerge attempting to overcome the disadvantages of previous designs while providing more features. Two popular database types, the Relational Database Management Systems and NoSQL databases are examined. The aim of this paper was to examine and compare two databases from these two database models and answer the question of whether one performs and scales better than the other.