thyroid-nodule

Deep Learning pipeline for thyroid nodule diagnosis in ultrasound. Benchmarks CNNs vs. Vision Transformers (YOLOv12, DINOv3) on a 7k+ dataset. Features a two-stage approach (Detection + Classification), achieving SOTA results with Foundation Models. Includes explainability maps and a GUI demo.

Generate impression from TI-RADS findings

TIRADS Calculator in Python

TensorFlow implementation of CNN architectures for thyroid nodule classification in ultrasound imaging

Large-scale thyroid ultrasound data analytics and EfficientNetB0-based AI classification using the TN5000 dataset, with leakage-aware preprocessing, statistical analysis, and a desktop research application.

Interactive simulation platform for thyroid nodule classification using ML, DL, and hybrid models. Built for education, visualization, and model evaluation on real ultrasound datasets.

LSTAC: An automated framework for segmentation, TIC analysis and peak intensity frame extraction of thyroid nodule CEUS videos.