Model description:

This model detects fire in all kinds of environments.

 

Number of classes: 1 (fire)

 

Metrics:

detections_count = 770, unique_truth_count = 344
class_id = 0, name = Fire, ap = 68.19%           (TP = 236, FP = 86)

for conf_thresh = 0.25, precision = 0.73, recall = 0.69, F1-score = 0.71
for conf_thresh = 0.25, TP = 236, FP = 86, FN = 108, average IoU = 53.37 %

IoU threshold = 50 %, used Area-Under-Curve for each unique Recall
mean average precision (mAP@0.50) = 0.681922, or 68.19 %

​

System requirements:

Inference time using CPU: 300 ms (on HP Laptop 15-DA0042NH (Processor: Intel(R) Core(TM) i7-8550U CPU))