CIFAR-10 Attention CNN and Brain Latent Diffusion Fine-Tuning
What it is
This is a two-part academic deep learning project consisting of a Selective-Kernel attention CNN for CIFAR-10 classification and a cohort-adaptive fine-tuning of a 3D MONAI Brain Latent Diffusion Model on OASIS-3 data. It includes a data pipeline for MRI synthesis and an evaluation suite to validate downstream Alzheimer's classification improvements.
The problem it solves
The scarcity of labeled medical imaging data for specific cohorts (like OASIS-3) limits the performance of downstream diagnostic classifiers; this project addresses this by generating synthetic data to augment training sets.
What you could use it for
- Training a Selective-Kernel attention CNN for image classification on CIFAR-10
- Fine-tuning a pretrained 3D brain latent diffusion model on a new cohort (OASIS-3)
- Evaluating the impact of synthetic MRI volumes on downstream Alzheimer's disease classification
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.