Projects

Ph.D Project Particulars

Name: Image-to-Image Transformation using Deep Learning.

Description: Currently I am working on Image-to-image transformation task, where the input image from one domain is transformed into the output image of another domain. It has many real world applications in image processing, computer graphics and computer vision like, image colorization, image super-resolution, image in-painting, image style transfer and Sketch-to-Photo synthesis, etc., A challenging issue in this task is to have a common framework that works for all the above mentioned applications. In this research, we aim to investigate and propose methodologies to perform inter conversion among various domains. Our research involves using Deep Learning methods such as Neural Networks to achieve the same.

M.Tech Project Particulars

Name : Iterative Weight Updated Kernel k-means for Multi-view Clustering.

Description : In multi-view clustering, for the data all available views are considered for clustering process that increases the clustering accuracy. But when we considering multiple views some of them are noise or degenerate views . so, their participation in clustering process may result in degradation of clustering accuracy. In order to overcome this problem we are proposed an iterative algorithm that solves this problem by giving weights to each view based on their contribution to the clustering process and updates these weights in each iteration with the help of an intra-cluster variance objective that is calculated at each iteration. We applied this algorithm for some of the multi-modal data sets and it show better performance than the available clustering algorithms. We implemented this in NetBeans IDE using JAVA.

B.Tech Project Particulars

Name : Information Retrieval form XML Documents.

Description : Information retrieval is fast becoming the dominant form of information access. Information retrieval focuses on organizing and storing data and retrieving useful information from it. With the advent of web, the XML has became a preferred standard for storing documents. This documents which are available on the Web can be structured as well as semi-structured. Traditionally, information retrieval was concentrating on retrieving the entire document that is relevant to the query. But with XML retrieval, the elements of the document at various levels of granularity can be retrieved instead of the document as a whole. We call this type of retrieval flexible retrieval. As documents can often be quite long , and searching the whole document is very complex, we retrieve parts of documents rather than the entire documents. We implemented this in C by Fedora using GCC Compiler and Libxml2 parser.