​

Dr. Suraiya Jabin along with her team is actively engaged in the implementation of three non-funded Collaboratory projects with Faculty of Dentistry in JMI, AIIMS, and the Department of Electrical Engineering, JMI.

​

Project 1: She with her team implemented an AI model for whole genome sequence (FASTQ/FQ) classification in 8 ESKAPEE categories. The highlight of this work is: ESKAPEE AI model can classify an input FASTQ sequence of size 3 Giga base pairs in less than 20 minutes towards 7 ESKAPEE and 1 non-ESKAPEE categories. This AI model is publicly deployed at http://115.241.23.53:8000/ for users or pathologists to upload their clinical raw sequence sample and get to know the type of ESKAPEE bacteria present in it in real-time. We are building on this research by developing another AI model, in collaboration with AIIMS, to predict antimicrobial resistance (AMR) phenotype in bacteria using whole genome sequencing (WGS) or next-generation sequencing (NGS) data. This model will identify whether a bacterial strain is resistant or susceptible to specific antibiotics.

​

Project 2: Another non-funded project is about the implementation of an AI model for mental task classification using EEG data collected in collaboration with Electrical Engineering department, JMI. An ethical clearance has been obtained from the Institutional Ethics Committee, JMI for enrolment of subjects and EEG data collection under novel data collection protocols.

​

Project 3: The third non-funded project work is under progress in collaboration with Faculty of Dentistry, JMI, and GE HealthCare. To support various activities of different departments such as implant planning, caries identification, disease diagnosis, etc., conducting CBCT scans is a regular task at Faculty of Dentistry, JMI. A total of at least 1000 CBCT volumes of the patients who reported or will report for various dental treatments, and are required to undergo a CBCT scan at the Faculty of Dentistry, JMI will be gathered retrospectively and prospectively. After anonymization and proper annotation, these will be used to build and train the AI model, mainly using techniques such as Vision Transformer and Self Supervised Learning to model the cognition of a dental radiology expert. This AI system will facilitate oral radiologists to engage in creative tasks by helping them facilitate their daily diagnoses. The proposed AI model once deployed, will be beneficial for reducing radiologists' workload and aiding dentists in treatment planning at JMI. A proper ethical clearance from the Institutional Ethics Committee (IEC), JMI has been obtained for this work.

All three projects mentioned above are under implementation and are non-funded currently.