Talk Summary:
AI tools for bioimaging have grown rapidly over the past five years. These tools have been developed by and for different research communities, including biologists, imaging scientists and computer vision researchers. Each community uses AI to address different problems, from understanding biological phenotypes to improving image acquisition and developing computational methods. This educational talk will first explore how these different needs shape the data requirements to applying AI tools effectively. It will then discuss emerging open data formats and frameworks that enable AI tools to efficiently access and analyse large imaging datasets. The talk will then highlight open-source and commercial AI tools across the imaging workflow, from image restoration and segmentation to image-based phenotyping. The examples will show how tool choice depends on the imaging problem, data requirements and the desired outputs. The talk will conclude by considering practical challenges in selecting, validating and integrating new AI tools, and opportunities to develop more generalisable approaches to AI-enabled bioimage analysis.

Presenter Biography:
Dr Akanksha Bhargava is a Lecturer and Early Career Academic Fellow in the School of Biomedical Engineering at UNSW Sydney, Australia. Her research investigates how imaging across scales and modalities can be integrated to understand biological structure and function in disease. She develops multimodal biomedical imaging and imaging data science approaches spanning micro-computed tomography, light-sheet microscopy, magnetic resonance imaging and hyperspectral imaging, with a focus on vascular and cellular dysfunction in cancer and neuroscience. Her PhD and postdoctoral research at Johns Hopkins University and Johns Hopkins Medicine, USA established her expertise in multiscale vascular imaging and image-based computational hemodynamic modeling, leading to the development of VascuViz, a multimodal vascular imaging framework for 3D mapping of vascular structure and function. This work also enabled the development of a three-dimensional in-silico angiogenesis atlas for breast cancer. At UNSW, she is extending this work through multimodal data integration and AI-enabled approaches to disease phenotyping and biomarker discovery. Her research has been recognised through the 2020 North American Microcirculatory Society Emerging Leaders Award and a 2025 UNSW Founders Early Career Academic Startup Fellowship supporting research translation in biomedical imaging.

Author
Akanksha Bhargava
University of New South Wales