Swarnendu Mukherjee

A Secure Scheme for Image Transformation

The dynamic growth in the field of Information communication has hiked the ease of Information transmission. But this type of real advancement in other sense has explored so many possibilities of information being snooped, at the time of communication in between the sender and the intended receiver. So, day by day Information Security is becoming an inseparable part of Computing and Communication. Various tools are available in order to address those issues and in them Steganography plays a significant role. In this paper, a heuristic approach for Information hiding in the form of multimedia objects or text using Steganography is proposed keeping in mind two considerations - Size and degree of security. Here, the Information is hidden behind an image in the form of an encoded matrix which is obtained from the bit matrix of the embedded object.

A Novel Approach for Determination of Optimal Number of Cluster

Image clustering and categorization is a means for high-level description of image content. In the field of content-based image retrieval (CBIR), the analysis of gray scale images has got very much importance because of its immense application starting from satellite images to medical images. But the analysis of an image with such number of gray shades becomes very complex, so, for simplicity we cluster the image into a lesser number of gray levels. Using K-Means clustering algorithm we can cluster an image to obtain segments. The main disadvantage of the k-means algorithm is that the number of clusters, K, must be supplied as a parameter. Again, this method does not specify the optimal cluster number. In this paper, we have provided a mathematical approach to determine the optimal cluster number of a clustered grayscale images. A simple index, based on the intra-cluster and inter-cluster distance measures has been proposed in this paper, which allows the number of clusters to be determined automatically.

A Novel Approach for Edge Detection of Images

Edge detection is a problem of fundamental importance in image analysis. Many approaches for edge detection have already revealed more are waiting to be. But edge detection using K-means algorithm is the most heuristic and unique approach. In this paper, we have proposed an algorithmic technique to detect the edge of any kind of true gray scale images considering the artificial features of the image as the feature set which is fed to K-Means algorithm for clustering the image and there to detect clearly the edges of the objects present in the considered image. The artificial features, which we have considered here, are mean, standard deviation, entropy and busyness of pixel intensity values.

A Secured Technique for Image Data Hiding

In this paper, a new technique for hiding the data of images has been proposed. This method is invented to hide an image file entirely with in another image file keeping two considerations in mind which are Size and Degree of Security. At the source, the image which is to be hidden (target image) is encoded at the end of another image (cover image). Double layer security of the hidden image can be achieved (over the untrusted network) by; firstly, the starting point of encoding the image data is depended on the size of the images and it is stored within the encoded image at the end of its header information as a cipher text.; secondly, the target image is hidden behind the cover image by following our encrypted image hiding technique.

Refine Crude Corpus for Opinion Mining

This paper is meant for a heuristic approach for the refinements of corpus based on regular expressions and its possible applications in the field of Opinion Mining. Corpus which is the plural form of ‘corpora’ is nothing but the collection of linguistic data. And here the proposed work is based on a corpus of reviews; more specifically product reviews. The reviews are in the html files which are easily available in popular review sites like Cnet.com. The revolution in information and technologies has given a new era in the development of language industries. The versatility in technological development, along with the translations available in different languages has lead to use of this corpus for specific machine learning mechanism as well as various automatic translation applications. But the prime objective of researchers as well as the naive users is to give a fast developing technique of machine learning systems that should be both exact and effective. Most of the time it becomes a very tedious job to create exact dataset for the work due to the crisis of accurate corpus regarding respective research work. And that is why; we have proposed an algorithm for creating a corpus for opinion mining research field.

Image Based Analysis of Tomato Leaves to Determine the Number of Petioles

The tomato (Lycopersicon esculentum) is an herbaceous, usually sprawling plant which belong to Solanaceae or nightshade family. Genetic evidence shows that the progenitors of tomatoes were herbaceous green plants with small green fruit. There are a great many (around 7500) tomato varieties grown for various purposes. Their identifications had been studied using various laboratory methods. The morphological and genetical characteristics were employed to classify different tomato cultivars. However, the presence of wide morphological varieties through evolution among various tomato cultivars made it more complex and difficult to classify them. Petioles plays a very crucial role in determining the characteristics of a tomato plant. The number of petioles present, their angle with the leaf stalk or their distance from the stalk represent genetical characteristics which differentiate various cultivars of tomato. This article proposed various methods to find the number of petioles present in a tomato leaf using an image analysis based approach.

A Novel Approach for Refinement of Corpus in the Field of Opinion Mining

In this paper, we have provided a heuristic approach for the refinements of corpus based on regular expressions and its possible applications in the field of Opinion Mining. The proposed work is based on a corpus of reviews. The crude corpus is the raw html files containing reviews. This html file is refined further for the ease of our work so that we can get only the required part from that page. The ultimate output yields the xml files which will precisely store the important parts of the review pages from that refined html page. And that is going to be fed to the further process of language processing for machine learning process in the field of Opinion Mining.

A Multi Layer Security Model for Text Messages

Of late in the field of Information Security, we have plenty of security tools which are made to protect the transmission of multimedia objects. But approaches for the security of text messages are comparatively less. In this paper, a security model is proposed which imposes the concept of secrecy over privacy for text messages. This model combines cryptography, steganography (taken as security layers) and along with that an extra layer of security has been imposed in between them. This newly introduced extra layer of security changes the format of normal encrypted message and the security layer followed by it embeds the encrypted message behind a multimedia cover object.

A Novel Approach for Designing Indian Regional Language Based Raw-Text Extractor and Unicode Font-Mapping Tool

Extracting specific information from a collection of documents is called Information Extraction (IE). In general, the information on the web is well structured in HTML or XML format. And the work of IE from structured documents (in HTML or XML), basically uses learning techniques for pattern matching in the content. In this paper, we have proposed a novel approach for interactive information extraction technique. Here, we have described how this approach enables any naive user to extract Indian regional language based document from a web document efficiently which is quite similar to a standard search engine. It is just similar to a pre-programmed information extraction engine.

A Heuristic Approach for Designing Regional Language Based Raw–Text Extractor and Unicode Font–Mapping Tool

Information Extraction (IE) is a type of information retrieval meant for extracting structured information. In general, the information on the web is well structured in HTML or XML format. And IE will be there to structure these documents, by using learning techniques for pattern matching in the content. A typical application of IE is to scan a set of documents written in a natural language and populate a database with the information extracted. In this paper, we have concentrated our research work to give a heuristic approach for interactive information extraction technique where the information is in Indian Regional Language. This enables any naive user to extract regional language (Indian) based document from a web document efficiently. It is just similar to a pre-programmed information extraction engine.