AI Leukemia Detection Tool

Upload a blood cell microscope image and get an AI-powered prediction with 87% accuracy. Built using EfficientNet-B3 and trained on the C-NMC 2019 dataset.

Medical Disclaimer: This tool is for educational and research purposes only. It should not be used for actual medical diagnosis. Always consult with qualified healthcare professionals for medical decisions.

Upload Blood Cell Image

Upload a microscope image of blood cells (preferably in .jpg, .png, or .bmp format)

Click to upload or drag and drop

Supports JPG, PNG, BMP formats

Image Preview

Analyzing image with AI model...

Analysis Results

Prediction

Confidence: 0%

How It Works

Model Architecture: EfficientNet-B3 with transfer learning
Training Data: C-NMC 2019 dataset with blood cell images
Accuracy: 87% on test dataset
Classes: Healthy (Hematopoietic) vs Leukemia (ALL)
Image Processing: Resized to 128x128 pixels with normalization

Technical Details

Framework: PyTorch with EfficientNet-PyTorch
Optimizer: AdaBelief with learning rate scheduling
Data Augmentation: Standard ImageNet normalization
Model Size: ~12MB (EfficientNet-B3)
Inference Time: ~200ms on CPU