FOOTBALL DETECTOR (YOLOV4)

On this website you can try the model with your own images and buy it if you are satisfied with its accuracy. 



 

Model description:

This model detects soccer balls.

 

Number of classes: 1 (football)

 

Metrics:

detections_count = 2420, unique_truth_count = 1102
class_id = 0, name = Football, ap = 92.60%           (TP = 994, FP = 160)

for conf_thresh = 0.25, precision = 0.86, recall = 0.90, F1-score = 0.88
for conf_thresh = 0.25, TP = 994, FP = 160, FN = 108, average IoU = 73.54 %

IoU threshold = 50 %, used Area-Under-Curve for each unique Recall
mean average precision (mAP@0.50) = 0.926039, or 92.60 %

System requirements:

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