This project aims to segment 4 common retinal lesions from Fundus Images.

Last update: Oct 10, 2021

Overview

To train a HEDNet model with conditional GAN to segment Hard Exudates using random seed 765, run python train_gan_ex.py --gan True --seed 765.
To train a HEDNet model with conditional GAN to segment Soft Exudates using random seed 765, run python train_gan_se.py --gan True --seed 765.
To train a HEDNet model with conditional GAN to segment Hemorrhages using random seed 765, run python train_gan_he.py --gan True --seed 765.
To train a HEDNet model with conditional GAN to segment Microaneurysms using random seed 765, run python train_gan_ma.py --gan True --seed 765.
When training HEDNet with cGAN, we apply all 3 preprocessing methods (Denoising + Contrast Enhancement + Brightness Balance).

To evaluate the model on the test set, run python evaluate_model.py --seed 765 --lesion 'MA' --model results/models_ma/model.pth.tar for evaluating a saved model checkpoint on MA under results/ using random seed 765. results/models_ma/model.pth.tar is the directory of the saved model checkpoint.

This project aims to segment 4 common retinal lesions from Fundus Images.

Related tags

Overview

Owner

Husam Nujaim

structured-generative-modeling

An Straight Dilated Network with Wavelet for image Deblurring

Pixel-wise segmentation on VOC2012 dataset using pytorch.

VACA: Designing Variational Graph Autoencoders for Interventional and Counterfactual Queries

An Agnostic Computer Vision Framework - Pluggable to any Training Library: Fastai, Pytorch-Lightning with more to come

Source code for CAST - Crisis Domain Adaptation Using Sequence-to-sequence Transformers (Accepted to ISCRAM 2021, CorePaper).

🎃 Core identification module of AI powerful point reading system platform.

Pytorch implementation of AREL

Synthesizing and manipulating 2048x1024 images with conditional GANs

Mail classification with tensorflow and MS Exchange Server (ham or spam).

Repository for "Space-Time Correspondence as a Contrastive Random Walk" (NeurIPS 2020)

A Planar RGB-D SLAM which utilizes Manhattan World structure to provide optimal camera pose trajectory while also providing a sparse reconstruction containing points, lines and planes, and a dense surfel-based reconstruction.

3D Pose Estimation for Vehicles

A short and easy PyTorch implementation of E(n) Equivariant Graph Neural Networks

High performance Cross-platform Inference-engine, you could run Anakin on x86-cpu,arm, nv-gpu, amd-gpu,bitmain and cambricon devices.

Adjust Decision Boundary for Class Imbalanced Learning

DeepMind Alchemy task environment: a meta-reinforcement learning benchmark

Implementation for "Manga Filling Style Conversion with Screentone Variational Autoencoder" (SIGGRAPH ASIA 2020 issue)

Airborne magnetic data of the Osborne Mine and Lightning Creek sill complex, Australia

Pytorch implementation of Depth-conditioned Dynamic Message Propagation forMonocular 3D Object Detection