Mrinal Kanti Ghose

Comparing some neural network models for software development effort prediction

Software development effort estimation is one of the most major activities in software project management. A number of models have been proposed to make software effort estimations but still no single model can predict the effort accurately. The need for accurate effort estimation in software industry is still a challenge. Artificial Neural Network models can be used for carrying out the effort estimations for developing a software project & this field of Soft Computing is suitable in effort estimations. The present paper is concerned with comparing the results of various artificial neural network models for predicting the software development effort estimation. The neural network models available in MATLAB neural network tools were used and the standard dataset as compiled by Lorenz et.al. was used in the present study. The results were analyzed using four different criterions MRE, MMRE, BRE and Pred. It is observed that the Generalised Regression Neural Network model provided better results.

Spatial analysis of environmental impacts of highway projects with special emphasis on mountainous area: an overview

Abstract Mountains are progressively being invaded by highways for development and defence purposes. Environmental impact assessment (EIA) of highway projects in mountainous areas creates a challenging environment for data collection and impact prediction. Geographic information systems (GIS)-based EIA, using appropriate spatial analysis methods can sufficiently reduce the challenges created by mountain environments. The present study includes a review of articles, including research papers, government reports and EIA reports cutting across the inter-disciplinary nature of the topic. The paper identifies the spatial analysis methods, models and salient features of conducting GIS-based EIA of highway projects in mountainous areas. It is observed that spatial analysis of impacts of highway projects on the environmental attributes, especially air, noise, water and socio-economy in mountainous areas are largely unexplored.

Multiple Image Encryption using Random Circular Grids and Recursive Image Hiding

visual secret sharing (VSS) scheme encrypts the secret information into various meaningless shares. These shares are distributed to the authorized participants and the secret information can be retrieved by any k out of n participants by stacking their respective shares on top of each other. This scheme uses HVS (Human Visual System) to decrypt the information, and thus no technical or financial investment is required. Moreover, it is a one-time pad technique, so decrypting the information by an attacker is almost impossible. This paper proposes an improved visual secret sharing technique in which we aim to build upon the random grid approach of visual cryptography and test the feasibility of Recursive Image Hiding to hide multiple secrets at varying levels of the grids generated. Since we are using circular random grids, it is even possible to hide multiple images in the same grids and obtain the secret images for different angles of rotation of the grids. The participants need to be in possession of both the shares, as well as the fixes angle of rotation for which the secret can be obtained, in order to decrypt the image. In case of recursive image hiding, numerous secrets are hidden recursively in the shares of the original images at each level. Shares carry information for the subsequent secrets as well, thus leading to increased capacity. Also, the limitation on the number of secrets that can be hidden can be overcome because for each grid, multiple secrets can be recursively hidden. Thus, not only will we be able to hide multiple images, but multiple grids as well which in turn carry the information for multiple images.

Selection of Defuzzification Method for Predicting the Early Stage Software Development Effort Using Mamdani FIS

The Soft Computing techniques presents an alternative to software development effort estimations and this paper is concerned with estimating the early stage effort estimations using student datasets. As we know, the output of fuzzy inference system depends on the selection of defuzzification method. This paper compares the outcomes of Mamdani Fuzzy Inference System (FIS) created using four different types of defuzzification methods namely centroid, bisector, Mean of Maximum (MOM) and Largest of Maximum (LOM). For this purpose a two inputs and one output Mamdani FIS was created using the above four defuzzification methods. The outputs (efforts) are evaluated using the standard performance evaluation metrics like Magnitude of Relative Error (MRE), Median of Magnitude of Relative Error (MMRE) etc. and it was found that the centroid method is the best and lom is worst among the defuzzification methods.

A novel approach for circular random grid with share authentication

Secrecy, Integrity and Authenticity are the major trust areas whenever information is subjected to third party intervention who can gain data access from vulnerable links in the system, communication channel or individual user identity. While addressing these issues, Cryptography turns out to be the trusted brand in the computer science fraternity and Visual Cryptography, as one among this, is a methodology where the secret recovery is done with human visual system without having to perform complex calculations. The proposed scheme complies with the methodology of secret sharing scheme where secret information is divided into various shares in meaningless form and is further recovered on overlapping printed transparencies with the shared information on it. Each of them is then validated for authenticity. An attempt has been made to use circular rings to embed the secret information with certain angular rotation and validation of the individual cipher shared in order to avoid cheating.

Geo-Statistical Modelling of Remote Sensing Data for Forest Carbon Estimation—A Case Study on Sikkim, Himalayas

The mountainous topography coupled with the high biodiversity throws up the range of challenges in remote sensing of forests for forest carbon assessment. The correlations between vegetation indices or grey-level co-occurrence matrix (GLCM) metrics and forest carbon cannot be reliably used in the estimation of forest carbon in biodiversity-rich mountainous topography of Sikkim Himalayas due to geometric, spectral and radiometric distortions. In this paper, an attempt has been made to apply the geo-statistical modelling of aboveground terrestrial vegetation carbon in Sikkim using products derived from remotely sensed satellite data, GIS and ground sampling data. The paper makes use of universal kriging as an interpolation method that makes use of semivariogram models for spatial autocorrelation to make a prediction of forest carbon at unsampled locations. The errors associated with prediction have been estimated. Remote sensing derived forest type and forest density maps have been used as cokriging parameters. Cross-validation technique has been used to evaluate the interpolation results. The analysis reveals that the total aboveground vegetation carbon stored in the forests of Sikkim is about 29.46 million tonnes. Importantly, the montane wet temperate forests contain the maximum forest carbon of 8.46 million tonnes, while the least carbon (1.03 million tonnes) is stored in the moist alpine scrub.

Why Green Computing and Green Information Sciences Have Potentialities in Academics and iSchools: Practice and Educational Perspectives

Green computing is responsible for the designing and developing the Computing Systems and IT infrastructure which are less energy consumed and eco friendly. The policies on environment, power management, consumption, recycling etc. have wider prospects in building and solid practice of Green computing and Green technology. Ultimately Green Computing is dedicated to the healthy and sophisticated Green Systems building. The integration of Green Computing with Informatics and Information Sciences have the potentiality to launched Green Informatics and Green Information Science respectively. The organization, association, government, educational institutes etc. have the healthy potentialities of introducing Green Computing and similar systems for healthy output and solid benefit. In the academics computers and similar devices become very common and thus the principles and theirs applications of Green Computing many ways would lead the sustainability. The iSchools are the consortium of information and computing related branches into one single academic wing and common in many countries. Thus they not only practice the Green Computing but also teach and engage the Green Computing and Green Information Sciences. The paper is described and reported Green Computing and allied branches their practice related issues and possibilities of the areas as academics subjects in Indian universities etc. The paper also highlighted the potentialities of iSchools regarding offering Green Computing and related branches including practicing Green Computing principles.

A Novel Educational Proposal and Strategies Toward Promoting Cloud Computing, Big Data, and Human–Computer Interaction in Engineering Colleges and Universities

Cloud computing is an important name in current Information Technology domain and responsible for the development in many perspectives. The advancement of information processing, management, delivery to software-related matter may be solved with cloud platforms. The virtualization of hardware, software, applications, etc., is in many ways dedicated for solid information infrastructure building and side by side solid information technology management. The related field of cloud computing is Big Data, which is responsible for managing large amount of data only rather software, hardware, etc. The application of Big Data is also called as Big Data Management. The complex and huge data generation results this domain and helps in data and information solutions in most of the organizations, institutions, associations, etc. Apart from these two, another important technology is human–computer interaction and partially related with the earlier mentioned technologies. It is worthy to note that these tools and technologies become common name in today’s world, but the educational opportunities are very much limited in India. The university and college level education as a full-fledged domain or specialization is more or less absent. This paper talks about cloud computing, Big Data, HCI as an overview including its current educational scenario with special focus on possible degrees, etc., in Indian context.

Steady Illumination Color Local Ternary Pattern as a Feature Extractor in Iris Authentication

The iris, considered as part of ocular biometrics, is an externally visible, yet protected organ whose unique epigenetic pattern remains stable throughout adult life. This characteristic makes it a very important modality for use as a biometric for authentication. When a subject wishes to be identified by iris recognition system, their eyes are first photographed, and then a template is created. This template is then compared with the other templates stored in a database until a correct match is found or it remains unidentified. Steady Illumination color Local Ternary Pattern (SIcLTP) has been used as a feature extractor in this paper to extract unique information from the color irises. The image matching is done using zero-mean sum of squared differences between the two equally sized images. The result shows that it can outperform the conventional local binary pattern as a texture descriptor.

IAABO-TCI: Improved Ant and Artificial Bees Swarm Computing Optimization techniques for content based image classification

The social insect metaphor for solving problems has become an emerging topic in the recent years emphasizing on stochastic construction practice, building the key probabilistically to optimize the solution related to any kind of a problem. As we are aware that image makes the human visualize the existence of entities in nature and helps individual to get the feel of the services without solely relying on the written messages. With the headway in image capturing devices, image data are being generated in high dimensions. However, processing the captured images plays a vital role that is gaining importance with its applicability in several filed of research. Numerous methods for processing of imagery have been developed nevertheless; existing methods still confines as the complexities of image increases with several coherent disadvantages such as greater amount of processing time and effort, prone to errors, inefficient outcome in terms of Accuracy Assessment parameters. Further, to classify the images into one of the several categories available in the database, typically make use of text associated with the image. The content-based image retrieval considering the visual information contained in the image itself is a challenging area to work on with. With this motivation in mind this paper focus on designing improved Swarm Computing methodology for content based image classification taking the full advantage of the solving power of Ant Colony Optimization (ACO) for Edge Detection, Support Vector Machines (SVM) as a base classifier, Discrete Wavelet Transform (DWT) for feature extraction and selection and Flower Pollination by Artificial Bee (FPAB) for optimization.