Skin Diseases Detection Using Deep Learning

30 Classes Based Skin Diseases Detection Using Deep Learning

Upload a dermoscopy or skin photo. The model ( EfficientNetB3 (phase 2) ) returns ranked class probabilities for 30 categories at 300×300 px, matching your training notebook pipeline.

EfficientNetB3 Architecture Visual

Optimized inference flow used by this model for 30-class skin disease prediction.

Input Layer

RGB image, resize to 300×300

→

EfficientNetB3 Backbone

Pretrained feature extractor

→

Global Avg Pooling

Compact deep features

→

Dense + Dropout

Regularized classifier head

→

Softmax Output

30 disease class probabilities

Drop an image here or click to browse

JPEG, PNG, WebP, BMP — max 16 MB