🧠 Segment brain tumors accurately using a 2D U-Net model in PyTorch with the BraTS 2020 dataset for efficient evaluation and visualization.

🧠 Implement U-Net in PyTorch for effective binary image segmentation, focusing on brain tumor detection with a complete pipeline from data prep to evaluation.

End-to-end UNet3D stack for volumetric medical segmentation: training utilities, FastAPI inference API, Streamlit review UI, Docker deployment, and a model card. Upload NIfTI volumes, get masks back, and visualize slices/IoU plots from experiments. Built for reproducible research and production-grade serving

Python programs to analyze GUV geometry (e.g., radius, volume), quantify protein binding in time-lapse confocal image data, and fit equilibrium binding curves to extract membrane affinity.

Generate consensus 3D cells segmentations by combining 2D cell segmentations from any combination of xy, xz, yz views, compatible with outputs of any 2D segmentation method.

[NeurIPS 2025] AutoSeg3D, online real-time 3D segmentation as instance tracking with long-short term query memory for embodied perception

SOTA 3D Medical Segmentation framework (UNET++, TransUNET , SwimTransUNET) in TensorFlow. Fully TPU-optimized with a 3-stage OHEM pipeline for the Medical Segmentation Decathlon (MSD).

Backend segmentation/classification library (using ONNX runtime) for Raidionics

Search for Protein 3D Coordinates in CryoET Tomograms

VAN-GAN: Vessel Segmentation Generative Adversarial Network. A tool to segment 3D vascular networks without paired training data.

MOOSE (Multi-organ objective segmentation) a data-centric AI solution that generates multilabel organ segmentations to facilitate systemic TB whole-person research.The pipeline is based on nn-UNet and has the capability to segment 120 unique tissue classes from a whole-body 18F-FDG PET/CT image.

Napari plugin for spatial RNA quantification

Winning Solution for Vesuvius Challenge First Title. Custom MiniUNETR 3D architecture for ink detection in carbonized scrolls.

The segmentation model for ECCV 2024 paper StyleCity

This is the official repository for Fast-nnUNet, a new fast model inference framework based on the nnUNet framework implementation.

A multi-organ tumour detection using deep learning. MSc First Class Honours project achieving 93.15% Dice score

Modified TensorFlow implementation of the 3D Mask R-CNN.

[NeurIPS 2024 Audio Imagination Workshop] Official implementation of the paper: 3D Audio-Visual Segmentation

🔥LogoSP in PyTorch (CVPR 2025)

3D U-Net model for volumetric semantic segmentation written in pytorch