Skip to main content

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

Lee - 2026

Mentor: Brian Lee | brian.lee2@nih.gov
Lab
Advanced Imaging and Microscopy Resource
NIBIB

Project 1: Artificial intelligence for image denoising and enhancement in teravoxel microscopy datasets (BESIP-BME)

AIM would like to submit two proposals for BESIP projects for this summer. I think these would be appropriate for the “BESIP-BME” program. We’re hoping to recruit one potential student for one of these projects.

Modern biological research generates vast volumes of microscopy imaging data as advanced imaging technologies enable high-resolution, high-throughput visualization of cells, tissues, and organisms. Computational methods involving artificial intelligence and computer vision are becoming increasingly essential for improving image quality, extending the boundaries of what can be visualized, and efficiently analyzing, interpreting, and extracting meaningful biological insights from this rapidly growing volume of data. The NIH's Advanced Imaging and Microscopy (AIM) resource is responsible for pioneering new biological imaging techniques and the computational methods that enable quantification and understanding of these images. AIM is seeking a BESIP student who will work with its staff scientists to develop new deep learning methodologies to improve image quality of its teravoxel 3D light sheet microscopy data. The student will investigate solutions in active areas of research such as intelligent fusion of multiple 3D microscope views and automated identification of salient markers to register massive image volumes. At the end of the internship, the intern will have had hands-on experience working with massive biological image datasets, will have made contributions towards a novel AI-based light sheet microscopy image enhancement method, and will have developed experience with approaching challenging research questions and developing a workflow to answer them.

Project 2: Development of deep learning models for few-shot object detection in light microscopy (BESIP-BME)

The NIH hosts some of the most wide-ranging types of biological research in the world, spanning many organisms, imaging modalities, and experimental contexts. Image segmentation is a critical task in many biological microscopy applications. However, it is impractical to develop and maintain bespoke artificial intelligence models tailored to every individual application, motivating the need for more general and adaptable approaches. The NIH's Advanced Imaging and Microscopy (AIM) resource is interested in recruiting a BESIP student to work alongside computer vision staff scientists to develop easily adaptable AI-based image segmentation and object detection models that can extend to new domains with few or no examples. The intern may be involved in the design and construction of foundation models trained on hundreds of terabytes of imaging data produced annually at AIM, as well as the design of new few-shot training and fine-tuning methods for image segmentation or object detection. At the end of the internship, the student will have had hands-on experience developing software for deep learning research and handling the machine learning workflow from data preparation to model evaluation, and will be exposed to real scientific questions relevant to the biological research community.