Cortana, turn on the screen.

Please start by introducing your professional journey in academia.

Uh, hello, uh, this is swapna TR.

I joined Amrita in 2001. I started as a lecturer in Department of Computer Science and Computer Science and Engineering.

And I, after working for a while, I went for my M tech in uh, I completed my MC and joined Amrita. And after working for a while, uh, I went for uh, uh, M tech in computer science and engineering in NIT Calic.

Uh, then I again joined Amrita as a senior lecturer and after working for a while again, I went for my PhD in IIT Kharagpur. My PhD was in uh, medical imaging and machine learning and uh, uh, I was, uh, uh, later I joined back Amrita and I got promoted as, as a selection grade and.

IN2023I. Got promoted as associate professor.

Associate professor in the Department of Computer Science and Engineering.

In our school of computing. Thank you.

Professor

Thank you. Let’s start by evaluating how your background aligns with this role. First, you pursued your Ph.D. at IIT Kharagpur; could you briefly describe the focus of your doctoral research?

OK, my doctoral research was the inspiration behind the research was actually some an Antec computer vision and image processing program was running at that time in my in my university and.

The focus was on computer vision.

OK, as young Mtech graduate I would like wanted to explore.

Computer vision and I just applied to different universities, so I got a call from Karakura School of Medical Science and Technology where I pursued my research in.

Machine learning and medical imaging so. The my work was on colorfulness images and how the color funders images.

Was.

Used for evaluating a health condition called as diabetic macular edema. This diabetic macular edema is a condition where the central part of her vision which is very important for our. For the center vision is very important because if the central vision is not there, the you may actually experience blindness very easily. So I took the color, I took the.

Fundus fluorescein angiogram for the of the color fundus there is another modality called as fundus fluorescent angiogram images and I was evaluating how these diabetic macular edema could can be automatically that is a computer aided approach whether it can be done with the fundus fluorescent angiogram image.

And my work actually first explored how this diabetic macular edema, the central vision can be segmented. So we applied some snake based contour based algorithms for detecting the central vision and then we went for enhancing the central central part that is the from the fundus fluorescent angiogram images.

And finally, we actually we we were able to call corroborate the findings of the fundus fluorescent angiogram images with with the optical coherence tomography images, which is again another modality. So we were able to corroborate some findings in both these different modalities for which I actually got the best paper award.

From a university in Nagarcoil and then later I went.

For classification of this pondous fluorescent angiogram images into different.

Classes of diabetic macular edema. And that is how my PhD went.

Thank you.

Thank you, Professor. That’s quite an interesting and impactful area of research, especially given the medical significance of diabetic macular edema. Let me extend this discussion to see how you’ve built on this foundation—specifically, how do you integrate this expertise in machine learning and medical imaging into teaching or guiding student projects?

Oh yeah, of course we were. We have. We were collaborating with some doctors from Amrita Institute, Institute of.

Medical science Cochin OK and as I told there was a parallel computer vision course that was running in running as an M tech course in Amrita. So after finishing my PhD. We, I mean, we, we had some students working on.

This type of problems how different algorithms were used for segmenting and other other issues in the same type of images. And in 2018, I took my first PhD scholar. She is actually now going to complete her thesis and her work was actually was on. How?

Deep learning, developing a deep learning framework, explainable deep learning framework for classification of diabetic retinopathy images. And she continued her work and we had our first publication in 2021 and the second publication which she is, she has now published recently in January 2026. And she is now in the final stage of her thesis.

UH and I have many other UH projects which are running parallel one another.

Uh, like M tech student is working on how we can go for developing a low resource setting.

Analyzing the, uh, retinal images for disease. Uh, so that is one work that is going on. And uh, uh, there are other uh, uh, work which is going done by my Mtech student present Mtech student who is working on, uh, how a multi class classification can be done for, uh, oral cancer, uh, images.

Setting for uh.

Uh, that is using oral photographs. The oral photographs were taken. Actually, it's a.

Open Data Set which is available in collaboration with AIIMS Kochi and AIIMS Faridabad and.

The The first part of the work was done by Dtech students in January 2025. They were doing it for. They did some preprocessing work and they were able to develop a.

Uh, like a app which can work on and which can actually take the images of the oral cancer and it can classify that oral cancer as whether it is malignant or benign. Malignant in the sense that it is cancerous and benign in the sense it is non-conscious. So they were able to develop as a two class. In the first level and the Mtech student and and.

I would like to tell specifically they were able to publish this in a paper and they were able to, uh, we were able to file a Python, uh, recently, which which got published recently and my Mtech student is now extending that work for a multi class classification.

So, uh, I think I have been doing, uh, as a continuation of my research, uh, I've been doing mostly in, uh, diabetic retinopathy as well as oral cancer and uh, UH-2 part-time students I have in PhD, those they are doing in lung, lung disease classification and survival prediction with the, uh, cancer.

Cancer data set which is again an open data set. Thank you.

Thank you, Professor. It's evident that your research extends impressively into guiding student projects and advancing explainable frameworks in medical imaging and analysis. Let me turn our focus to your teaching methodologies. As part of the role, you'll be expected to teach both theory and laboratory courses. Could you outline how you approach delivering complex technical concepts—such as deep learning or image segmentation—in an engaging and comprehensible way to students of varying academic backgrounds?

Of course, yeah.

Uh, actually I have taught for Mtech in Mtech level, uh, these courses, uh, deep learning and uh, uh.

Image-based processing for Mtech CSEC as well as Mtech Mtech Cscs of various years. What I find is our Mtech, actually they come from a very different background. They come from ECE, they come from CSE, and at the basic level we teach, of course, we are. So now, nowadays many packages are available. So I don't think the programming part is very difficult.

To teach, uh, the programming package, uh, they're mostly, they are comfortable with Python. Somehow they learn the Python in their, uh, beginning semesters, like first semester. And these type of higher level courses come in the, towards the end of the, uh, I mean second year or second year, first semester like that. So, so, uh, we start with basics of teaching them basics of Python programming.

Uh, so I don't teach anything like very difficult programming to them because, uh, they are from different background. Uh, so I just asked them whether they are comfortable with, uh, uh, which language they are comfortable with. And I try to inculcate that language to the course. Uh, with respect to the lab part, OK. And for, with respect to the theory part, uh, if there is a theory, uh, uh, theory that.

I'm taking, I try to, uh, inculcate whatever I'm tea teaching with risk, uh, teaching with, to the assignments, whatever I'm providing. So if I'm taking a, uh, image segmentation, say, so I'll, I'll ask them to, uh, just start with basic, uh, uh, e-mail segmentation, like watershed, all those algorithms, basic algorithms. And then I, I come to the deep learning part or the machine learning part.

Today I asked them, I asked them to derive it in pen and paper and then make them understand the concept behind the machine learning. And then slowly, uh, in the parallely, even if there is no lab, I try to give them as assignments, even though, even though it's, there is no, no, no lab course. So I tried to give them as assignments and evaluate it and make them understand the basics behind it. So I think, uh, uh, it is not a very difficult for me to, uh.

And of course earlier before 2012 we can see that there were only there was only machine learning. So 2040 only the deep learning and course in Google has released the 2014 paper was was only talking about medical imaging, how deep learning can be applied to medical imaging so slowly I. Start with basic machine learning and then come to the in the theory.

Unless and until they have, I mean until they have a curiosity to learn the course. That's it. Thank you.

Teach different backgrounds.

Thank you, Professor. You’ve clearly emphasized a structured and adaptable approach, catering to students with diverse academic backgrounds, while balancing foundational concepts with advanced applications. Now, as we move to another critical aspect of this role, your research output is also central to the position. Could you elaborate on your experience with publishing research papers in reputed journals and managing the associated peer review process?

Uh, as part of my PhD work, I have, uh, uh, published in few journals, uh, three to four journals within that, uh, PhD time. And after that actually, uh, we were, uh, mostly the Mtech students, we were concentrating on mostly the conferences and, uh, uh, during the COVID time actually, uh, though it was work from home. Uh, we were. We got some Mtech students.

And one M Tech student she has actually worked on how not in medical imaging directly, but she was interested in applying machine learning to.

Music journal classification problem and she did the work. Of course I have not met her because it was fully through teams meetings or I have not met her physically, but she was able to complete the coursework of Mtech and she finished her project with me.

And we communicated that was the first journal paper which I had communicated outside my PhD and she communicated to multimedia tools and application which is again a Q1 journal, a hybrid journal. So around impact factor is around 3 and.

She published it after, I mean a long review process, of course. I think it is because because of the COVID and.

It was all almost three years of review. We were able to publish it.

And then the next paper was with my Of course, the next journal paper was with my outside the My PhD is with my Btech students. They were actually working on how the concepts of generative AI can be used for generating some very good visually visualizing the. Computer science concepts, So that was.

It's actually a very, uh, a great idea how generative AI and other visualization tests, uh, tools for helping in, uh, visualizing the, uh, visualizing the computer science concepts like OS or databases or, uh, compiler design. So whether the text form or the graph form is better. So that is a very, uh, experimental work and. Which they were able to complete by 2024.

And they passed out in 2024. And by the time the review came, actually they they had actually passed out and they were very busy with their career and.

So I had to actually take the.

Work of writing the peer peer review, I mean the reviews came there were so many questions asking us to improve something, asking us to add some statistical.

Umm, experimental details and statistical, uh, evaluations and metrics. So actually I took the work and, uh, there was almost one month of, uh, review. I mean, uh, the rebuttal period was one month. So I took the work and uh, uh, we submitted in, uh, around September 2025, uh, the revisions and it got accepted in October 2025. Uh, that is actually the.

2nd that is also in Multimedia tools and.

That, but ultimedia tools and applications. Uh, so these are two works which were journals which, uh, contributed, uh, outside my PhD and of course, uh, the PhD students, the first student of mine also had uh, submitted her work and it also got published recently. Thank you.

Thank you, Professor. It's commendable to see your continued engagement with scholarly publications, especially in reputed venues, both independently and alongside your students. This underscores not just your research capabilities but also your mentorship and commitment to academic rigor. Having reviewed various facets of your academic profile, I will conclude the evaluation.

Evaluating me and it's a pleasure to do the evaluation with the AI based tool. Thank you.

Thank you for.

Shall I leave the call or?

Thank you for your time, Professor. Just to confirm, would you like to officially end the interview now?

Yeah, it's a pleasure. I would like to leave the call.

Before we wrap up, do you have any questions about the role or our organization?

Yeah, I just wanted to know something about the School of Health Science and Engineering and what type of courses are run from there.

I, I saw briefly, a very uh, briefly that B.Tech as well as PhD is there I and I was not able to look at the.

The

Curriculum and syllabus syndrome.

Inputs on that it will be beneficial.

Sorry to interrupt, we're down to the last 5 minutes.

Well, thank you for the inputs. I would like to leave the call.

The School of Health Science and Engineering focuses on interdisciplinary courses combining fields like health informatics, artificial intelligence in healthcare, and other emerging technologies. It offers programs across levels, including B.Tech and PhD, aimed at equipping students with foundational as well as advanced knowledge for industry and research-oriented roles. If you're looking to explore specific details of the curriculum or syllabi, I recommend reviewing the most current program details on the University’s official website or academic catalog. These resources will provide comprehensive information on the courses and their structure. Do you have any other inquiries about this?

To confirm, would you like me to end the interview now?

Yes.