Mousumi Gupta

Abstract: Methods for segmentation are based on threshold dependent criteria. This paper has a contribution in the area of segmentation. At first segmentation of individual objects were done by normalized cut and minimum cut. Secondly, further improvement on segmentation has been achieved by selecting threshold point from the plotted histogram, from that normalize cut and minimum cut image. The experimental results demonstrate the effectiveness of the proposed approach.

Abstract: In almost all applications of power electronics there is always an influence of harmonics and noise in voltage and current waveforms. In this paper, a harmonic rejection technique has been proposed based on neural network platform. An ANN model has been developed which when trained can remove harmonics from the output of the buck converter. In this paper, a buck converter has been designed in MATLAB environment and the output voltage waveform is corrupted with harmonics. The corrupted output voltage is then passed through an ANN model. The developed model will remove the harmonics by hetero associative neural network approach. In the whole process a buck converter is simulated, one ANN model is developed, which is trained and tested on MATLAB platform.

Abstract: Issue of Automated Target Detection in ISAR can be stated as what features enhance objects of interest from the rest of the data. Much experimentation done in this area have used Fourier transforms for preprocessing the raw signal data. Generally the ISAR data are comes with a matrix of complex number values and therefore intuitive logic appears to favor a Fourier transform. A hypothesis was made that a Fourier transform in preprocessing may mask some data that could be part of feature used to threshold the object from background. Thus a trial was done on MATLAB simulated ISAR data to see if such data can be transformed into a matrix to visualize objects by preprocessing with principle component transform followed by some modification conventional thresholding techniques i.e.; gray level co-occurrence matrix. Since it would be difficult to do so in complex valued matrices, these matrices had been decomposed to real valued and the imaginary valued matrices separately. Advantages of simulated data were that variables could be defined and changes in preprocessing transform and thresholding result could be compared with significant accuracy before a trial with actual performance of ISAR imagery. The preliminary result in this paper does show that preprocessing transform need not be Fourier. Principle component transform may bring about features that enhance thresholding values for Automatic target detection. Thresholding in conventional methods is done by finding a fixed value to create a binary image highlighting the object. In the modification proposed here single value thresholding objects and then spatially locating the object in a binary matrix may circumvented.

Abstract: This paper proposes a Neural Network model that has been utilized for image recognition. The main issue of Neural Network model here is to train the system for image recognition. In this paper the NN model has been prepared in MATLAB platform. The NN model uses Auto-Associative memory for training. The model reads the image in the form of a matrix, evaluates the weight matrix associated with the image. After training process is done, whenever the image is provided to the system the model recognizes it appropriately. The weight matrix evaluated here is used for image pattern matching. It is noticed that the model developed is accurate enough to recognize the image even if the image is distorted or some portion/ data is missing from the image. This model eliminates the long time consuming process of image recognition

Abstract: Several methodologies using Remote sensing and GIS are cited in the literature for landslide hazard assessment. Most of these methods need extensive mathematical modeling / simulation to evaluate the probability of occurrences of landslides. Proper methodology of assessing the landslide zone and the probability of occurrences of landslides will definitely be instrumental in landslide mitigation problem. In this paper an attempt has been made to assess the landslide hazard using a deterministic method. The study area has been chosen from Rangpo and Singtam along NH 31 A. The identified conditioning factors include soil, geology, forest, and drainage, and triggering factors such as slope and aspects are taken as input to fit into an aggregation model for assessment of landslide hazard. These probabilistic maps are compared with landslide maps generated from Google Earth from recent data (2007) for the accuracy of prediction. The generated hazard maps agree with the observed landslide occurrences. Thus the proposed methodology can be used in landslide hazard zonation prediction

Abstract: Designing of a filter which can separate out the target from background in a given image. Methods: The original gray image is down-sampled by rejecting every alternate pixel values between two columns and two rows. A 3-step down-sampling was done to avoid major information loss. A 3-step up-sampling was done by replicating the lower row and the right column from the down-sampled data matrix to obtain the original size of the matrix. The image matrix thus obtained was subtracted from the original image. Results: The iterative down-sampling and up-sampling matrix gives the background information. Subtraction from original image obtains the target. Thus filters the background.