Informatics Publishing Limited

P. Sivaranjani ¹, A. Senthil Kumar ²

Affiliations
1 Department of Electronics and Communication Engineering, Kongu Engineering College, Perundurai – 638052, Tamil Nadu, IN
2 Department of Electrical and Electronics Engineering, Dr. Mahalingam College of Engineering and Technology, Pollachi - 642003, Tamil Nadu, IN

Abstract

Background: Floor planning is important step in physical design automation of VLSI circuits, because it gives an early feedback on the architectural design. It is the process of finding the position of the module such that no two modules overlap with each other. Methods: In order to have an efficient floor plan, the total area occupied by the modules should be minimum. So, non-slicing floor plan is used to find an optimal floor plan layout. To represent non-slicing floor plan, a number of representations are proposed. Findings: To encompass billions of transistors in an Integrated Circuit (IC), 3Dimensional (3D) IC is preferred instead of 2D. In this paper, a novel 3Dimensional (3D) non-slicing floor planning representation called Modified Corner List (MCL) algorithm is proposed and properties of MCL algorithm is derived. The proposed algorithm is illustrated using Microelectronics Centre of North Carolina (MCNC) benchmark circuits and simulation results shows that it is very effective for 3D floor plan representation. Improvements: The proposed algorithm works well for small number of modules. As the number of module increases, computational time taken by the algorithm also increases. The above problem can be solved by applying heuristic algorithm in association with MCL strategy to find near optimal placement in reduced run time.

Keywords

MCNC Benchmark Circuits, Modified Corner List, 3D Non-Slicing Floor Plan, VLSI

Full Text

   

S. Maheswaran ¹, M. Ramya ¹, P. Priyadharshini ¹, P. Sivaranjani ¹

Affiliations
1 Department of ECE, Kongu Engineering College, Perundurai - 638052, Tamil Nadu, IN

Abstract

Objective: This paper discusses about the development of hardware system which can be used to chase the birds from the farm by using an embedded technology. Methods: The system is able to detect and track the pest birds from the real time video frame using Kalman filter. The Kalman filter is a predictive filter, that dependent on space techniques and recursive algorithms using image processing tools in Matlab environment. The various states of operation of a dynamic system are estimated using this algorithm. Applications/Improvements: The proposed system has been simulated in Matlab. It can be extended to a complete the bird frightening system by incorporating the video processing algorithm with hardware such as camera, linux based embedded board and ultrasound generator. The video obtained from the camera is processed using the ARM7 Raspberry Pi board based on linux OS. This system detects the birds from the video and a high frequency ultrasound is generated and directed towards the pest birds to drive them away from the farm.

Keywords

Kalman Filter, Linux Based Embedded Board, Ultrasound Generator.

Full Text

   

P. Keerthana ¹, P. SivaRanjani ², G. Sree Gayathri Devi ³, S. Shanmugapriya ³, G.H. Sindhuja ³

Affiliations
1 Assistant Professor, Department of ECE, Kongu Engineering College, Perundurai, IN
2 Professor, Department of ECE, Kongu Engineering College, Perundurai, IN
3 Student, Department of ECE, Kongu Engineering College, Perundurai, IN

Abstract

In today’s modern world, diseases are increasing day by day. Amid the different types of diseases, cancer has turned out to be a deadliest disease in human beings. According to the latest medical statistics, Cancer is accountable for high mortality rate. Skin cancer is usually identified beginning with preliminary screening by clinicians and followed possibly by dermoscopic investigation, a biopsy and histopathological inspection. Differentiating and predicting different types of skin lesions on inspection of images is one of the most challenging and difficult task because of the minute variations in the appearance of skin lesions. The images of different lesions seem to be identical and it is very difficult to manually differentiate them. Hence, deep convolutional neural networks are now emerging as a solution to address this problem. The main objective of this work is to review the indexed papers that addresses the issues of automatic classification of skin lesions. This paper summarizes about the different types of datasets available, the type of deep learning models used for training and the parameters used for performance measurements.

Keywords

Skin lesion classification, Deep Learning, Data Augmentation, Artifacts, Transfer Learning.

Full Text