Model Performance Metrics
Detailed accuracy analysis across both AI classification models — Skin lesion classifier and Monkeypox detection.
74.14%
Skin Lesion
95.0%
Monkeypox
2
Models
Overall Accuracy
Convolutional Neural Network trained on 10,000 dermatoscopic images across 7 lesion categories.
Total Classes
7
Lesion categories
Architecture
CNN
.h5 / Keras format
Best Class
vasc
100% accuracy
Weakest Class
mel
49.3% accuracy
Per-Class Accuracy
Skin Lesion Classifier
Per-Class Breakdown
7 classes| Code | Condition | Accuracy |
|---|
Overall Accuracy
TFLite quantized binary classifier trained to detect Monkeypox virus infection from skin lesion images with high sensitivity.
Detection Confidence
Correct vs error rate
Sensitivity
95%
True positive rate
Specificity
95%
True negative rate
Per-Class Accuracy
Binary classification breakdown
Class Breakdown
Monkeypox
Monkeypox Virus Infection
Not Monkeypox
Other Skin Conditions
Model Technical Details
Input Size
256×256
pixels
Format
TFLite
Quantized
Threshold
0.50
binary cutoff
Output Type
Binary
2 classes
