AI for Science

The transformative potential of AI and Machine Learning in enabling and accelerating scientific discovery is immense. AI algorithms and models empower domain experts in the fields of medicine, natural sciences, and engineering to explore vast swathes of parameter search spaces in order to test their hypotheses and optimize their designs. The IMST Lab has collaborated with clinicians in the visual sciences for over a decade and half designing corrective software for AMD, developed classical and dee learning based AI/ML imaging approaches to diagnose, treat, and manage eye diseases including AMD, Diabetic Retinopathy, Uveitis, infectious diseases such as Toxoplasmosis, and inherited retinal disease (IRD). IMST Lab has also been engaged with material scientists, biologists, and engineers in developing AI/ML models and methods to analyze 2D materials and biofilms based on deep learning based image analyses of microscopy and spectroscopy images whose quality is improved using deep learning based super-resolution methods. We have successfully completed federal (NSF, VA) and state (NRI) supported projects in this area.

Data Driven Material Discovery Center for Bioengineering Innovation (2023 - Present): NSF Award #2324925

UNO: M. Subramaniam/Parvathi Chundi (Co-PI), SDSM: R. Winter (PI), B. Jasthi (Co-PI), USD: C. Lushbough, E. Gniempieba (Co-PIs), MSU: M. Fields (Co-PI)

Description Microbes attached to surfaces, commonly known as biofilms, represent multi-million dollar challenges and opportunities in municipal water, marine, manufacturing and oil and gas sectors and a range of other engineering and medical applications. The study of biofilms at the cellular level and the study of materials at the atomic level generate extremely large amounts of rich data. To mine this data and establish connections between biofilm growth and material properties, this Research Infrastructure Improvement Track-2 Focused EPSCoR Collaborations (RII Track-2 FEC) award will form a new collaboration between South Dakota School of Mines and Technology, Montana State University, the University of Nebraska - Omaha and the University of South Dakota to develop Big Data Analytic Tools. This team will develop the Biofilms Data and Information Discovery system (Biofilm-DIDs) to collect and combine these large data sets using artificial intelligence to analyze and predict gene responses and biofilm characteristics influenced by surface properties. By accomplishing this goal the team intends to rapidly accelerate the pace of discovery of new materials to control and leverage biofilm growth. This project will provide education, training and workforce development opportunities for a diverse cohort of junior faculty and post-doctoral researchers and graduate, undergraduate and high-school teachers and students.

The primary objective of this project is to develop Big Data Analytic Tools for understanding rules of life in biofilms on technologically relevant materials modified with an emerging class of single-atom thick, two-dimensional (2D) materials. This will be accomplished by developing the Data Driven Material Discovery (DDMD) Center for Bioengineering Innovation, which will coalesce diverse infrastructure in bioscience, computer science, and material science from South Dakota School of Mines & Technology, Montana State University, the University of Nebraska-Omaha and the University of South Dakota to develop the unique Biofilm-DID system. The DDMD Center will focus on the development of novel interdisciplinary approaches and data analytics to track biofilm phenotypes on 2D materials, coupled with -omics analyses of sulfate-reducing biofilm phenotypes to discover rules of biofilm assembly and organization governed by atomic-scale material surface features. The DDMD Center?s areas of research will include: big data mining, machine learning, and predictive modeling; 2D materials for biological applications; and biofilm composition and diversity. The Biofilm-DIDs will be developed, calibrated and validated to provide a scientific platform for interrogating biological mechanisms in response to nano-scale properties. This platform will be leveraged to understand how the substrate crystallographic orientations and point defects in coatings affect gene expression, signaling pathways, metabolites, and structure formation controlling stress resistance, extracellular electron transfer, and biocorrosion mechanisms of biofilms. The DDMD center infrastructure will offer a series of education, training, and workforce development opportunities in data analytics and informatics approaches customized to material and biofilm sciences. 

Outcomes Project is ongoing.

Publications

• P Thakur, V Gopalakrishnan, P Saxena, M Subramaniam, KM Goh, B. Peyton, M. Fields, R.K. Sani,Influence of Copper on Oleidesulovibrio alaskensis G20 Biofilm Formation, Microorganisms 12 (9), 1747., 2024. (PubMed: PMC11434458)

• V Bommanapally, D Abeyrathna, P Chundi, M Subramaniam, Super resolution-based methodology for self-supervised segmentation of microscopy images, Frontiers in Microbiology Vol. 15, pp 122850, 2024. (PubMed: PMC10963421)

• Md. R. Islam, M. Subramaniam, P. Huang, Image-based Deep Learning for Smart Digital Twins: a Review, arXiv:2401.02523v1, January 2024.

• D. Abeyrathna, Label Efficient Learning for Multi-Label Classification With Self-Supervision, Doctoral Dissertation, University of Nebraska at Omaha, 2024.

• J. LS Zylla, E. Gnimpieba, A. B. Bomgni, R. K. Sani, M. Subramaniam, C. Lushbough, R. Winter, V. Gadhamshetty, P. Chundi, Convergence research and training in computational bioengineering: a case study on AI/ML driven biofilm-material interaction discovery, Biomedical Engineering Education, Vol. 4, pp:283-294, 2023. (PubMed: PMC10503862)

• V. Bommanapally, M. Subramaniam, S. Talluri, V. Gadhamshetty and J. Jones, Self-Supervised Scribble-based Segmentation of Single Cells in Biofilms, IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Istanbul, Turkiye, 2023.

• Md H. Rahman, V. Bommanapally, D. Abeyrathna, Md Ashaduzzman, M. Tripathi, M. Zahan, M. Subramaniam, V. Gadhamshetty, Machine Learning-Assisted Optical Detection of Multilayer Hexagonal Boron Nitride for Enhanced Characterization and Analysis,, IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Istanbul, Turkiye, 2023. 

•  D. Abeyrathna, M. Subramaniam, P. Chundi, An Overview of Machine Learning, Machine Learning in 2D Materials Science, pp: 33-56, CRC Press 2023.

• P. Chundi, V. Bommanapally, V. Gadhamshetty, The Future of Data Science in Materials Science,, Machine Learning in 2D Materials Science, pp: 217-232, CRC Press 2023.

• Md Ashaduzzaman, M. Subramaniam, Self-Supervised Learning-Based Classification of Scanning Electron Microscope Images of Biofilms,  Machine Learning in 2D Materials Science, pp: 109-129, CRC Press 2023.

• Md H. Rahman, M. Tripathi, A. Dalton, M. Subramaniam, S. Talluri, B. K. Jasthi, V. Gadhamshetty, Machine Learning-Guided Optical and Raman Spectroscopy Characterization of 2D Materials, Machine Learning in 2D Materials Science, pp: 163-175, CRC Press 2023.

• F. Shahjin, M. Patel, M. Hasan, J. D. Cohen, F. Islam, Md Ashaduzzaman, M. U. Nayan, M. Subramaniam, Y. Zhou, I. Andreu, H. E. Gendelman, B. D. Kevadiya, Development of a porous layer-by-layer microsphere with branched aliphatic hydrocarbon porogens,, Nanomedicine: Nanotechnology, Biology and Medicine, Elsevier, 2023. (PubMed:  PMC10460474) 

• An AI-Based Approach for Detecting Cells and Microbial Byproducts in Low Volume Scanning Electron Microscope Images of Biofilms, Frontiers in Microbiology, Vol. 13. pp: 4867, 2022 (PubMed:  PMC9751328)

• A. D. Chakravarthy, D. Abeyrathna, M. Subramaniam, P. Chundi, V. Gadhamshetty, Semantic image segmentation using scant pixel annotations, Machine Learning and Knowledge Extraction, 4(3), pp: 621-640, 2022.

• V. Bommanapally, Md Ashaduzzman, M. Subramaniam, S. Talluri, V. Gadhamshetty, Leveraging Weak annotations for Deep learning tasks on Biofilm Images, IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Las Vegas, 2022.

• Md Ashaduzzman, V. Bommanapally, M. Subramaniam, P. Chundi, J. Kalimuthu, S. Talluri, V. Gadhamshetty, Using deep learning super-resolution for improved segmentation of sem biofilm images, IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Las Vegas, 2022.

• P. Huang, E. Shakya, M. Song, M. Subramaniam, Biomdse: A multimodal deep learning-based search engine framework for biofilm documents classifications,  IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Las Vegas, 2022.

• V. Bommanapally, Md Ashaduzzman, M. Malshe, P. Chundi, M. Subramaniam, Self-supervised learning approach to detect corrosion products in biofilm images, IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Online, 2021.

AI-driven Image Analytics for Visual Science (2012 - Present): Supported by NRI, VA.

UNO: M. Subramaniam/P. Chundi, UNMC: E. Margalit, Stanford: Q. D. Nguyen, Cincinnati: R. Sisk

Description  The IMST Lab's research in AI-driven image analytics in the area of Visual Sciences and Ophthalamology are focused in two broad areas -- the analyses of various imaging data to aid clinicians in assessing retinal health, detect structural changes, guide treament decisions, and manage eye conditions. The IMST Lab has been engaged in projects involving clinicians internationally from Turkey, England, Brazil, Phillipines, India and the US developing software, AI-models, and platforms. The IMST Lab has ongoing collaborations in the US with Cincinnati, Stanford, and Utah medical schools. The Lab has also extensively collaborated with the UNMC visual sciences department.

Imaging A number of imaging modalities and technologies are employed by clinicians to assess eye health. The IMST lab has experience with several commonly used imaging modalities such as the Fundus Photography, Optical Coherence Tomography (OCT), Fluorescein Angiography (FA),  as well as high-resolution Adaptive Optics images, and developed image processing algorithms, AI/ML models that can perform image classification, segmentation, and object counting tasks to assist clinicians.

Publications

• O. Elaraby, A. Dasgupta, D. Abeyrathna, M. Subramaniam, Q. D. Ngyen, Objective AI Quantification of Longitudinal Goldman Visual Field (GVF) Changes Secondary to Autoimmune and Inherited Retinal Diseases, Second UCLA/American Uveitis Society International Workshop, 2024.

• V. Bommanapally, A. Akhavanrezayat, Q. D. Nguyen, M. Subramaniam, COINS: Counting Cones Using Inpainting Based Self-Supervised Learning, In Proc. of 46th Annual International Conference of the IEEE Engineering and  Biology Society, 2024.

• A. Akhavanrezayat, V. Bommanapally, D. Abeyrathna, M. S. Halim, C. Or, I. Karaca, G. Uludag, N. Yavari, V. Bazojoo, A. Mobasserian, Y. Shin, M. Hasanreisoglu, P. Chundi, Q. D. Nguyen, M. Subramaniam, A novel objective method to detect the foveal center point in the rtx1TM device using artificial intelligence, Investigative Ophthalmology and Visual Science (IOVS), 64(8), 2023 (Technology under Copyright 2024).

• M. Hasanreisoglu, M. S. Halim, A. D. Chakravarthy, M. S. Ormaechea, G. Uludag, M. Hassan, H. B. Ozdemir, P. C. Ozdal, D. Colombero, M. N Rudzinski, B. A. Schlaen, Y. Sepah, P. Chundi, M. Subramaniam, Q. D. Nguyen, Ocular Toxoplasmosis Lesion Detection on Fundus Photograph using a Deep Learning Model, Investigative Ophthalmology and Visual Science (IOVS), 61(7), 2020.

•  D. Abeyrathna, M. Subramaniam, P. Chundi, M. Hasanreisoglu, M. S. Halim, P. C. Ozdal, Q. D. Nguyen, Directed fine tuning using feature clustering for instance segmentation of toxoplasmosis fundus images, In Proc. of 20th IEEE Intl. Conference on Bioinformatics and Bioengineering (BIBE), 2020.

• A. D. Chakravarthy, M. Subramaniam, P. Chundi, S. Ragi, V. Gadhamshetty, Thrifty Annotation Generation Approach for Semantic Segmentation of Biofilms, In Proc. of 20th IEEE Intl. Conference on Bioinformatics and Bioengineering (BIBE), 2020.

• A. D. Chakravarthy, D. Abeyrathna, M. Subramaniam, P. Chundi, M. S. Halim, M. Hasanreisoglu, Y. J. Sepah, Q. D. Nguyen, An Approach Towards Automatic Detection of Toxoplasmosis using Fundus Images, In Proc. of 19th IEEE International Conference on Bioinformatics and Bioengineering, (BIBE), 2019.

Visual Distortion Detection and Correction

• M. Hassan, A.D. Chakravarthy, M. Subramaniam, P. Chundi, M.A. Sadiq, M. S. Halim, R. Afridi, A.N.T. Tran, Y. Sepah, D. V. Do, Q. D. Nguyen,  Correction of perceived visual distortions using a software application and correlation to age-related macular degeneration, Therapeutic Advances in Ophthalmology, Vol. 12, 2020.  (PubMed: PMC7235661)

• A. D. Chakravarthy, M. Subramaniam, P. Chundi, M. Hassan, Q. D. Nguyen, Towards Automated Distortion and Health Correlation for Age-Related Macular Degeneration,  In Proc. of 17th IEEE International Conference on Bioinformatics and Bioengineering, (BIBE), 2017.

• P. Chundi, M. Subramaniam, K. Sabet, E. Margalit,  Analyzing retinal optical coherence tomography images using differential spatial pyramid matching, In Proc. of 16th IEEE  International Conference on Bioinformatics and Bioengineering (BIBE), 2016.

• S. Go, P. Chundi, M. Subramaniam, E. Margalit, Analyzing OCT images of age-related macular degeneration patients to identify spatial health  correlations. in Proc. of 37th Annual International Conference of the IEEE Engineering in Medicine and Biology, 2015.