BrainAid
A lightweight hybrid deep learning model for four-class classification of brain tumor MRI images. It combines CNN and Vision Transformer (ViT) architectures and incorporates an explainable AI module, and can be deployed on edge devices such as Raspberry Pi.
Timeline (last 90 days)
The BrainAid system uses a hybrid CNN and Vision Transformer (ViT) architecture for a four-class brain tumor classification task (glioma, meningioma, pituitary tumor, no tumor).
BrainAid achieved 98% precision, a 98% F1 score and 98% accuracy on public datasets.
BrainAid's performance metrics surpass those of standard models such as VGG16, Inception-V3 and ResNet-50.
BrainAid has been deployed to run on a Raspberry Pi, verifying its feasibility in low-resource environments.
BrainAid integrates an explainable AI (XAI) module that visualizes the model's decision process and helps clinicians understand which image regions the model focuses on.
What BrainAid demonstrates is experimental results on public datasets; real clinical deployment still requires cross-institutional, prospective external validation and reporting of performance differences across subgroups.
BrainAid is positioned as a front-line screening and decision-support step, not a replacement for radiologists.
All Facts (7)
BrainAid is positioned as a front-line screening and decision-support step, not a replacement for radiologists.
50%UnverifiedBrainAid achieved 98% precision, a 98% F1 score and 98% accuracy on public datasets.
50%UnverifiedBrainAid's performance metrics surpass those of standard models such as VGG16, Inception-V3 and ResNet-50.
50%UnverifiedThe BrainAid system uses a hybrid CNN and Vision Transformer (ViT) architecture for a four-class brain tumor classification task (glioma, meningioma, pituitary tumor, no tumor).
50%UnverifiedBrainAid integrates an explainable AI (XAI) module that visualizes the model's decision process and helps clinicians understand which image regions the model focuses on.
50%UnverifiedWhat BrainAid demonstrates is experimental results on public datasets; real clinical deployment still requires cross-institutional, prospective external validation and reporting of performance differences across subgroups.
50%UnverifiedBrainAid has been deployed to run on a Raspberry Pi, verifying its feasibility in low-resource environments.
50%