Prabira Kumar Sethy

A security enhanced approach for video Steganography using K-Means clustering and direct mapping

Communication security has taken vital role with the advancement in digital communication. The universal use of internet for communication has increased the attacks to users. The security of information is the present issue related to privacy and safety during storage and communication. Cryptography and Steganography are two popular ways of sending essential information in a confidential way. Cryptography is the method of converting plain text into cipher text but in Steganography messages are converted into an encrypted format using a key and then this cipher text is hidden into an image, audio or video file as per users choice. The information-hiding process in a steganographic system starts by identifying a cover mediums redundant bits (those that can be modified without destroying that mediums integrity). The embedding process starts with creating a stego medium by replacing these redundant bits with data from the hidden message. In this paper, we present a novel approach to resolve the remained problems such as robustness and capacity of image and video Steganography. In the proposed algorithm, message bits are clustered and grouped together using K-Means clustering and then the clustered message is embedded inside the cover medium by using direct mapping which result increase the robustness and capacity of the cover medium. The robustness specially would be increased against those intended attacks which try to reveal the hidden message and also some unintended attacks like noise addition as well.

Network structure based protocols for Wireless Sensor Networks

The Wireless Sensor Network (WSN) is a wireless network consisting of ten to thousand small nodes with sensing, computing and wireless communication capabilities. WSN are generally used to monitor activities and report events, such as fire, overheating etc. in a specific area or environment. It routs data back to the Base Station (BS). Data transmission is usually a multi-hop from node to node towards the BS. Sensor nodes are limited in power, computational and communication bandwidth. Primary goal of researchers is to find the energy efficient routing protocol. This study highlights the recent routing protocols for sensor networks and presents a classification for the various approaches pursued. The three main categories explored in this paper are data-centric, hierarchical and location-based. Each routing protocol is described and discussed under the appropriate category with advantages and limitations. The paper concludes with issues open for research.

Fault Diagnosis in Wireless Sensor Network using Timed Automata

ABSTRACT An important problem in distributed systems that are subject to component failures is the distributed diagnosis problem. In distributed diagnosis, each working node must maintain correct information about the status (working or failed) of each component in the system. In this paper we consider the problem of identifying faulty (crashed) nodes in a wireless sensor network and used timed automata for representation. A fault diagnosis protocol specifically designed for wireless sensor networks is introduced and analyzed using finite automata theory. The protocol is proved to be optimal and energy efficient under certain assumptions. In this paper, we propose a diagnosis algorithm on the basis of diagnosability definitions and theoretical studies developed for timed and hybrid automata. The proposed algorithm has been simulated by using MATLAB and the diagnosis parameters such as diagnosis latency and message complexity

Horticultural Approach for Detection, Categorization and Enumeration of On Plant Oval Shaped Fruits

The basic and primary step of any image processing approach, which classifies the similar areas in the image and helps in further analysis, is Segmentation. This paper reports a segmentation algorithm for automatic singulation, categorization and enumeration of on-plant oval shaped fruits for satisfying the purpose of automatic yield estimation. The algorithm is based on thresholding of color channels that are derived from specific color spaces. Thresholding of RGB color space has been used in the process of singulation and thresholding of YCbCr color space has been used in the process of categorization. In the process of enumeration, edge detection and dilation operations have been used. Results obtained were satisfactory basing upon various performance metrics.

Detection and Categorization of OnPlant Prunus Persica: A Novel Approach

The approach presented in this paper is for onplant detection of Prunus persica (peach fruit) and their classification, basing upon their maturity level. Image processing techniques used for this purpose are color-based segmentation and thresholding. The proposed approach effectively distinguishes and categorizes the sample images of mature and immature images of peaches. The efficiency of the algorithm can be estimated from its accuracy, precision and recall value, which were found out to be 0.60, 0.60 and 0.74 respectively. The algorithm is time efficient and can result in significant reduction of cost and human labor, if implemented in an automated system for performing onplant peach detection and categorization.

Detection and Identification of Downy Mildew Diseased Leaf of Cucurbita Pepo (Pumpkin) Using Digital Image Processing

Since India is a largest international producer of pumpkins, downy mildew disease of pumpkin diminishes the foremost fiscal progression toward the field of agriculture. Year after year from the very beginning to till the date, it has been noticed that Cucurbita pepo harvesting for our eatable purposes and remains indispensable harvest plant forever. Due to its edible demand, it is highly produced and available on a large scale for its commercial purposes. These plants have many diseases and infections but downy mildew is one of the most and common appearing disease among all diseases. This paper herewith proposed an original approach toward segmentation algorithm, i.e., K-means clustering for automatic detection and principal component analysis (PCA) for identifying the diseased leaves which help to indorse the early disease detection. The trial significances infer that the planned scheme might help to identify and distinguish the mentioned disease satisfactorily and successfully from the leaves of pumpkin.

Measurement of Disease Severity of Rice Crop Using Machine Learning and Computational Intelligence

This study was conducted to develop a prototype which computes the severity of diseases appears in the rice crop using machine learning and computational intelligence. The symptoms of rice crop diseases imply the seriousness of the disease and suggest choosing the best approach to dealing with the disease. Most of the diseases in rice crop appear as a spot on the leaves. It is also needful to diagnose the disease properly and on-time to avoid the great harm of the rice crop. The treatment of rice crop diseases by applying disproportionate pesticides increases the cost and environmental pollution. So the use of pesticides must be minimized. This can be actualizing by targeting the diseased area, with the appropriate quantity and concentration of pesticide by estimating disease severity. This paper introduces Fuzzy Logic with K-Means segmentation technique to compute the degree of disease severity of leaves in rice crop. The proposed method estimated to give up to about 86.35% of accuracy.

Sensing Technology for Detecting Insects in a Paddy Crop Field Using Optical Sensor

This paper proposed a system which is to detect insects in a paddy crop field. Today we are living in the twenty-first century where computer vision is playing important role in human life. Computer vision provides image acquisition, processing, analyzing, and understanding images and, in general, high quality image from the real world in order to produce numerical or symbolic information, in the forms of decisions. It provides not only comfort but also efficiency and time saving. Today satellites are used as computer vision technology; by analyzation of the satellite images, it gives the information to the user. But this is only applicable for scientific level research laboratory because the cost of this type of devices is very high and not suitable for using in a farm field. So here we design a system, which detects insects in a farm filed and population estimation of insects in a farm field. The objectives of this paper are to control pests in a farm field and a healthy crop yielding for increased food production.

Application of Soft Computing in Crop Management

Indian agriculture is overwhelmed by numerous complications; some of them are usual, and some others are artificial like small and fragmented land-holdings, seeds, manures, crop selection, crop planning, fertilizers and biocides, irrigation, lack of mechanization, soil erosion, agricultural marketing, inadequate storage facilities, and so on. With the progression of different and specific outfits for the viability test of crop management are essential for providing reliable data observing to the performance of crop management. Valuable practical data can be collected by utilizing fuzzy logic-based scheme, in contrast with the intrinsic objectivity for collecting the data in gradual progression without any flaw. By dint of subject expertise and with the knowledge of scientific derivation, the approach should inspire to every corners of the country and management of cropping schemes. This paper analyzes the application of soft computing techniques in crop management in the field of farming and organic engineering is manifested. Upcoming progress and implementation using soft computing in the arena of farming and organic work to be think about.

Identification and counting of mature apple fruit based on BP feed forward neural network

Classification of fruits is an onerous and tedious task because of countless number of fruits. The traditional approach for detection and classification of fruit and its maturity level is based on the naked eye observation by the experts, which is both time consuming and causes eye fatigue. Advance techniques in image processing and machine learning helps to automatic classify and count the fruits, accurately, quickly and non-destructively. A method to automatic detect and classify apple fruit maturity level, whether it is mature or immature based on its color features has been proposed. Images of the apple are resized and Image Processing Techniques are applied for the extraction of apple color components (R, G, B). Artificial Neural Network is used as a classifier to identify and count the mature and immature applesusingcolor components. The proposed model has an accuracy of 98.1%, when all the three attributes are used as an input.