MedEvalKit: A Unified Medical Evaluation Framework

An implementation of the paper Brain2Qwerty that translates brain EEG data into text for reading people's brains. There was no code so we made an implementation

MICCAI 2024: Welcome to the official repository of D-MASTER: Mask Annealed Transformer for Unsupervised Domain Adaptation in Breast Cancer Detection from Mammograms. This repository hosts the source code, pre-trained model weights, and benchmark dataset RSNA-BSD1K, supporting research in cross-domain breast cancer detection.

Skin Diseases Detection System

This deep learning model detects eye diseases using CNNs. Trained on an image dataset, it predicts conditions with high accuracy. Ideal for AI-driven medical diagnosis!

A deep learning model for classifying eye diseases using Convolutional Neural Networks (CNNs). Built with TensorFlow and Keras, trained on a medical eye dataset for accurate multi-class classification and analysis.

🔬 Aplikasi deteksi penyakit kulit menggunakan AI dan Computer Vision - Lomba Digiwar #1: Digital Application Challenge - VISIONARY: Unleashing Innovation through Computer Vision.

Production-ready RAG chatbot for heart disease information using FastAPI, PostgreSQL + pgvector (Neon), and React + Vite.

A Telegram bot that classifies photos of skin lesions using a fine-tuned Vision Transformer (ViT-B/16) from HuggingFace, then fires a text-only label to Google Gemini 2.0 Flash for a plain-English breakdown: what the condition is, what signs to watch for, treatment options, and a one-line severity verdict. Seven HAM10000 diagnostic classes.

Deep learning model using VGG16 to classify chest X-rays into COVID-19, Pneumonia, TB, and Normal. Deployed via a local web app with real-time predictions.

This project applies machine learning to predict heart disease using clinical data. It covers data preprocessing, model building, and performance evaluation, aiming to support early diagnosis and healthcare decision-making through data-driven insights and AI-based prediction techniques.

PyTorch-based pipeline includes data preprocessing, model inference, and performance evaluation with standard metrics (Dice score, Hausdorff distance). The repository provides tools for visualizing segmentation results and comparing MedSAM-2's performance against baseline models, offering insights into adapting foundation models for medical imaging

This repository presents an efficient approach for fine-tuning large language models for the medical domain using 4-bit quantization and LoRA techniques.

ai assisted emergency triage and hospital routing to reduce er overload and improve response time.

An Agentic RAG system using LangGraph Designed to answer medical questions accurately and contextually

This project builds a deep-learning-based heartbeat sound classification system using MFCC features and multiple models including CNN, BiLSTM, and a Hybrid CNN–BiLSTM architecture. The system detects and classifies heart sounds into normal, murmur, and artifact categories, supporting early cardiac abnormality detection.

Parallel processing models: Gemini Pro, Gemini Flash, and GPT-5 Mini. A Chief AI Officer (GPT-5) evaluates all outputs, selects the strongest elements from each model, and synthesizes them into a single, unified result. Turn complex medical records into understandable health data

Deep learning model for brain cancer detection using CNN architecture. Trained on multi-class MRI data with ImageDataGenerator and optimized with Adam

A Noise-Resilient Hybrid Imputation-Ensemble (NR-HIE) framework designed to bridge the generalizability gap in medical AI. Utilizing a triple-stream imputation strategy and stacked generalization, this model achieved 81.62% accuracy on external validation data, ensuring robust and medically safe diabetes prediction