The department of Computer Science and Engineering is pioneer in research. It has diverse set of trained faculties in various domains of research. Many faculties in the department are pursuing their research in various domains which are hot trends. The department have wide variety of research areas like :

•    Data Science
•    Big Data
•    Machine Learning
•    Digital Image Processing
•    Natural Language Processing
•    Networks & Security
•    Artificial Intelligence
•    Data Mining
•    Information Security
•    Computer Vision
•    Digital Signal Processing
•    Medical Image Processing
•    Software Engineering


Research Projects :


Finished Research Paper





Mrs. Mausumi Goswami

Term Frequency and Inverse Document Frequency method for better Text Representation


Type : Working Paper



Every aspect of society is influenced by digitization as the information age infiltrates society. It has caused exponential growth of digital data. Different forms of origin of such data can be categorized as unstructured text on the world wide web, sensor data, digital images, videos, sound, result of scientific experiments and user profiles for marketing. Large text datasets and the high dimensionality associated with representation and processing of text datasets is quite challenging to retrieve useful information from them. This is caused by the various characteristics associated with natural languages and a major concern in text mining. In this research, a systematic study is conducted, in which few different document representation methods for text are used. A comparative study of various text representation techniques is implemented. It is found that term frequency and inverse document frequency can be used to enhance representation of text present in documents.  Text clustering is one of the central problems in text mining and information retrieval area. In the area of information retrieval document clustering plays a very important role. The reason behind it is investigated. In this research clustering techniques are explored to evaluate text representation. Two main techniques used are k-means and fuzzy clustering for text documents. Next, validation of the clustering result is done using silhouette coefficient. The selected feature retains original physical meaning and provides a better understanding for the data and learning process.


Dr K Balachandran

Software Model by Data mining classification and feature selection for Respiratory Cancer diseases


Type: Minor Research Project




Data mining is the process of extracting knowledge from the data and finding hidden relationship present in the data. Respiratory cancer disease is one of the leading cause of death in the modern world. ‘Surveillance, Epidemiology, and End Results Program’ maintains a record of cancer patients. The data acquired is used for data mining study. The data taken for the study belongs to different time period and different data format. Hence pre-processing of the data has been carried out to bring the data to a common format. Nearly a million records have been taken for this study. This research work focused on predicting the class labels (type of respiratory disease) based on the Levenberg-Marquardt algorithm which is based on Gauss-Newton and Steepest Descent algorithms. This algorithm works based on the back propagation approach.



Mr. Vinai George

Formal Verification Of Requirements Design Using Uml And Colored Petri-Nets


Type: Working Paper




Verifying requirements during software development process modelling systems behavior and analyzing system correctness is checked through Colored Petri-Nets. This validation ensures that the software process remains with established parameters under anticipated conditions. Since UML does not provide formal approach for specification and has weak support for validation, CPN tools with state space validation is included as a part of the requirements engineering process.






On Going Research Paper


Mr.Michel Moses

A  Framework for Summarization of surveillance videos based on Semantic feature selection and dimensionality reduction.


Type: Minor Research Project




            Every day enormous video data made by surveillance cameras are being collated and stored. These video sets not only takes up a huge space for accumulation of data but takes a long time to identify the content that is being searched. The efficient way of storing video data is to remove high-degree redundancies.   This calls for techniques to concise the space and time utilized by using Automatic Summarization of Surveillance videos (ASOSV). The three properties that are to be kept in mind are Un-supervision, Efficiency and Scalability. In this work major focus is given to store only important frames that are non-redundant and are highly unique to describe data. Dimensionality reduction based on Sub Space analysis will help in reducing the multi-dimensional data into a low-dimensional data to enable faster feature extraction and summarization.


Dr. Samiksha Shukla

Secure Collaborative Computation framework using Graph Theory

Type: Monograph


The advancement in internet usage and availability of affordable access is enhancing the collaboration. This collaboration could be advantageous when secure joint computation can be performed on the individual data. As this collaboration could be between mutually untrusted players or competitors, we need to ensure security of individual involved. In this research work we will be using concept of graph theory to design the framework for Secure Collaborative Computation.


Mrs. Gokulapriya

Study on pattern generation-a data mining perspective


Type: Working Paper



Human activity understanding includes activity recognition and activity pattern discovery. This focuses on accurate detection of the human activities based on a predefined activity model. Activity recognition builds a conceptual model first, and then implements the model to develop a decision support system. This work involves initial study regarding different pattern generation methodologies. The study helps more about finding some unknown patterns directly from low-level data without any predefined models or assumptions. Activity pattern discovery builds a system first and then analyzes the data to discover activity patterns. Monitoring human activity and finding abnormality in their activities used by many field like medical applications, security systems etc. Basically it helps and support in decision making systems.

Mr Bijeesh

Implementing Level-set theory based algorithm for human ear image segmentation on Graphical Processing Units.


Type: Working Paper



Image processing is a fast growing area of research which has its roots in both mathematics and computer science. Researchers all over the world have been working hard on reducing the time taken by various image processing tasks. In applications such as medical imaging, the execution time is a critical factor. Complex image processing can be computationally very high time-consuming even for recent computer hardware. In contrast to CPU, Graphical Processing Unit (GPU) is very efficient in parallel processing. Though the original purpose of GPUs was to provide a graphic output to the users, their computing capability, has led to the use of GPUs for general purpose computing which is known as General Purpose GPU (GPGPU). GPUs have an architecture that supports massive parallelization, as they can execute thousands of threads concurrently. Image processing is an obvious choice for parallelization because pixels can be mapped directly to threads and a lot of data is shared between pixels. This work proposes to implement a level-set theory based algorithm for human ear image segmentation on GPU. Human identification based on ear biometric has been found to be very effective. However, segmenting the ear image from the background is the toughest challenge in implementing an ear biometric system. Level-set theory based algorithms are effective in image segmentation. This paper aims at parallelizing the segmentation algorithm and implementing it on GPU using the Compute Unified Device Architecture (CUDA) framework.



Research Scholars :



Title of Research

Area of Specialization


Sujatha A K

Anomaly Detection in Online Social Media

Machine Learning


Gokulapriya R

Behavior modelling approach for identity managment system using integrated data mining techniques

Data Mining


Kukatlapalli Pradeep Kumar

Secure Provenance based Communication using Encryption Techniques

Information Security


Boppuru Rudra Prathap

Spatio Temporal Analysis on Social Media data

Machine Learning


Mausumi Goswami

Evaluation, identification of coherent patterns using computational intelligence technique

Data Science


Michael Moses T

Semantic Summarization of Surveillance Videos through Performance and Learning framework.

Computer Vision


Chinthakunta Manjunath

Web vulnerabilties

Web Technology


Mukesh Kamath

Joint Compression and Encryption of Multimedia Data

Graphics and Multimedia


Shridhar Venkatanarasimhan 

Direction Finding Algorithms for Electro-Magnetic Wave Radiators and Acoustic Emitters, JNTU Hyderabad

Digital Signal Processing



Classification of pap smear cervical cells

Medical Image Processing


Mahesh D S 

Efficient Bandwidth Scheduling in wimax networks using evolutionary computation technique



Praveen Naik 

Development of Effective Cost Estimation for Web Application

Software Engineering


Rohini G 

A Framework for Multimedia Streaming Using Cloud Techniques, Anna University ,Chennai.

Cloud Computing


Research facilities :