🏥🎾 Universal Cross-Domain Vision Model

Classifies images across medical (X-ray pathologies) and sports domains using an ensemble of BiomedCLIP, ViT-B/16, ResNet-50, and EfficientNet-B0 backbones with fine-tuned multi-modal attention fusion.

Medical classes: Normal, Pneumonia, COVID-19, Tuberculosis, Cardiomegaly, Rib Fracture, Lung Mass, Pleural Effusion Sports classes: Running, Jumping, Swimming, Cycling, Tennis, Football

Upload any image to get started.