PADDY DISEASE DETECTION
Detection studio
Compare your baseline, initial proposed and final proposed models on the same image.
Input image
JPG / JPEG · PNG · WebPImages are processed on your device and are not uploaded to a detection server.
CONTROLLED COMPARISON
The same input.
The same settings.
Each model receives the same image, confidence threshold and input size. Explore bounding boxes, predicted classes and confidence side by side.
Follow the same leaf.
Detection results
Real model predictions appear here after comparison.
RECORDED EXPERIMENTAL RESULTS
Performance, with perspective.
Detection quality and computational complexity across the evaluated models.
DETECTION PERFORMANCE
Quality across models
Percent · higher is better · axis starts at zero
COMPUTATIONAL COMPLEXITY
Cost across models
Lower is lighter · axis starts at zero
PERFORMANCE-COMPLEXITY TRADE-OFF
Where each model stands
mAP@0.5:0.95 score difference
56.08% → 55.83%
mAP@0.5 score difference
80.69% → 81.67%
parameters
reported GFLOPs
CLASS-LEVEL DETECTION PERFORMANCE
Every class, three perspectives.
Compare your baseline, initial proposed and final proposed models for each class.
Test-set performance · percent · higher is better · all bars use a 0-100% scale.
Exact measurements
| Model | Precision | Recall | mAP@0.5 | mAP@0.5:0.95 | Parameters | GFLOPs |
|---|