Inspiration

Our team at Nero Imaging is dedicated to aiding neurobiologists by providing advanced tools to measure and analyze anomalies and deformations in the neurological system.

What it does

IRIS Vector Search leverages deep learning to perform precise segmentation of mouse brain images, identifying tiny regions and abnormalities within the neuro system. This tool is critical for researchers focusing on specific areas such as the hippocampus, enabling the calculation and measurement of anomalies that are vital for advancing neurological health.

How we built it

We developed the ISis text response and notification system to facilitate communication with crucial service agencies like hospitals and emergency responders. This system leverages the IRIS Vector Search technology, built upon a robust tech stack. We utilized the Python Flask framework for server-side operations, incorporating HTML, CSS, and JavaScript for a dynamic and responsive frontend design. SQLite3 serves as the backbone for our database management, ensuring efficient data handling and storage.

To address the challenges posed by the large medical image datasets, which typically range from 25 to 35 GB per volume, we innovated by transforming these massive 3D images into more manageable 2D patch images. This approach significantly reduces the computational load while maintaining the necessary detail for high-quality analysis and segmentation. This method allows for more efficient processing and analysis, essential for advancing our understanding of neurogenetic conditions and enhancing our diagnostic capabilities.

Challenges we ran into

One of the main challenges was managing the enormous size of the original 3D brain volume data, which is difficult to process on standard desktop computers.

Accomplishments that we're proud of

At Nero Imaging, we're proud of our progress with the IRIS Vector Search, particularly our ability to segment the hippocampus in mouse brain images. Despite challenges in prediction accuracy, the foundation we've laid is crucial for future breakthroughs. Our integration of complex data handling and computational technologies is another highlight. Although we haven't yet achieved predictive success, the upcoming hackathon represents a promising opportunity to reach this goal. We are excited by the potential to transform our capabilities and make significant contributions to neurobiological research.

What we learned

Our journey underscored the importance of robust computational resources and the power of AI in medical imaging. We've gained insights into the complexities of brain structure and the precision required for effective vector search.

What's next for ISis-InterSystem Intelligent Solutions

Looking forward, we are focused on enhancing our segmentation capabilities, and potentially expanding our technology to other areas of medical research. Our next goal is to refine our system for broader applications, including more complex segmentation tasks that could further revolutionize neurobiological studies and clinical applications.

Built With

Share this project:

Updates