Personal Website

Tri-Annot: A Three-Stage Consensus Framework for Automated Nodule Annotation

NoMAISI (Nodule-Oriented Medical AI for Synthetic Imaging), a generative framework built on foundational backbones with flow-based diffusion and ControlNet conditioning.

CancerRisk is a repository for developing and evaluating cancer risk prediction models.

PiNS — Point-driven Nodule Segmentation Weakly supervised 3D lung nodule segmentation, radiomics feature extraction, and patch generation from bounding-box or point annotations. Includes ready-to-u…

CaNA (Context-Aware Nodule Augmentation) is a medical imaging toolkit that leverages organ and body segmentation masks as contextual guidance to augment lung nodule segmentation masks.

Developed a pseudo-labeling AI framework combining ensemble models, calibration, and unsupervised alignment to curate large clinical CT datasets.

Developed a pseudo-labeling AI framework combining ensemble models, calibration, and unsupervised alignment to curate large clinical CT datasets. [

This study presents the development and validation of AI models for both nodule detection and cancer classification tasks. This benchmarking across multiple datasets establishes the DLCSD as a reli…

A rule-based algorithm enabled the automatic extraction of disease labels from tens of thousands of radiology reports. These weak labels were used to create deep learning models to classify multipl…

SYN-LUNGS: Towards Simulating Lung Nodules with Anatomy-Informed Digital Twins for AI Training

Will contain the Intro to ML Duke ECE580 TA tasks,

Using COVID-19 as a case study, this virtual imaging trial study verified that radiology AI models often suffer from a reproducibility crisis.

My Personal Websie

To evaluate the performance of each regularization method (cutout, mixup, and self-supervised rotation predictor), we apply it to the CIFAR-10 dataset using a deep residual network with a depth of …