Unusual activity detection is the process of identifying and detecting the activities which are different from actual or well-defined set of activities and attract human attention. We have developed an application to automatically detect unusual activities. Generally, humans follow a sequence in their actions. Based upon these sequences of activities, we will classify them into the usual and unusual activities. A lot of human resources are required to monitor the vast number of surveillance cameras installed in public and private places. So, it is highly recommended to have an automatic algorithm which can detect unusual activities and trigger action.

The objective of the project is to develop better toxicity assessment features, methods and algorithms along with better data structures to handle the big data. To create new methods for assessing chemical toxicity which will have the potential to improve the procedure followed by scientists to evaluate environmental chemicals and develop new medicines.To develop a framework to quickly and efficiently test certain chemical compounds for their probable chances to disrupt processes in the human body.To develop a stand-alone and/or web-based application(s) that helps the researchers and research community to predict the toxicity of the newly discovered chemical compound.

The research project has been conceptualized with a goal to use data science effectively for aetiology research in India. The research aims to use machine learning for prediction of gene expression levels and knowledge discovery methods, and come up with a dashboard to handle different tasks related to data science. Specific objectives include development of tools to quantitate epigenomic signatures and mapping the localized nature of protein modification, and correlation of epigenomic signatures and gene expression using the developed tools. We will also develop a framework of reporting to various stakeholders regarding the status of different desirable parameters for policy making and decision making.

As the internet continues to grow and infiltrate every aspect of an individual's life, the security and crime related aspects of the cyber-world also show up their ugly face. Cyber bullying, Frauds, Cyber crime and other nefarious online activities regularly create headlines. Data science approaches and technologies can become the first line of defence which shall combine together text mining, machine learning and statistical approaches to provide security threat prediction, detection and prevention at an early stage. The objectives of the research project are to effectively detect and combat a variety of cyber crime issues such as cyber bullying, cyber predation, social media trolling, social media spam, etc. The research is envisioned to provide tools and software for use by an individual or a group which shall help them remain safe from such threats while performing their regular online activities.

“Communication in the sky” is a trend because the increment of Unmanned Aerial Vehicles (UAVs) use in wireless communications. UAVs have diverse applications in civil and military domains. Swarm of UAV system is able to finish the operations more reliably than a single UAV. UAV system have rapidly changing topology due to high mobility devices. Currently, Mobile adhoc networks (MANETs) routing is used for communication in UAV networks, and the standards for communication system are yet to be developed. UAV system streaming information needs protocol with high bandwidth, high mobility, varying link stability and high energy consumption compare to adhoc networks. It leads to abruptly breaking communication in between UAV-to-UAV and UAV-to-ground. Hence to overcome link breakage, stability in the links is required for the duration of next route establishment. We explore the link stability issue in UAV communication systems. Link stability is required for reliable communication.

Developing improved knowledge discovery techniques based on Machine Learning and Soft Computing tools specially Multi-objective optimization algorithms, fuzzy theory, Support Vector Machine etc. and applying the techniques for analyzing biological data such as transcriptomics data (e.g. microarray mRNA/microRNA expression data, RNA-Seq data) in order to unveil potential biological processes, genetic markers and reveal transcriptional regulatory mechanisms involved in the development of certain cell types such as cardiomyocytes, etc., progression of diseases such as breast cancer, etc. Identifying such genetic markers may provide new insights into therapeutic targets in diseases.

The BIONEST project aims to develop a reliable infrastructure for establishing the identity of new born infants and managing it to facilitate a robust biometric identification. The project aims to identify the reliable biometric characteristic for establishing the identity of the infant, and employ the video captured using smartphone in a non-intrusive and contactless manner to determine the liveness by estimating the vitals of the infant. The third component of this project is aimed at developing the privacy preserving techniques for infant biometrics. This is a joint project proposal between IIT-BHU, Bennett University and NTNU Norway.