Evaluation Metrics

1 bbox_mAP（bounding box mean Average Precision）

"bbox" refers to the bounding box. In the object detection task, the main task of the model is to predict the position and category of each object in the image. Usually, the position of an object is represented by a rectangular bounding box. "mAP" is the abbreviation of Mean Average Precision. "bbox_mAP" is the mean average precision of the bounding box. It comprehensively considers the detection accuracy of the object detection model under different categories and different IoU (Intersection over Union) thresholds, and is a key indicator for measuring the performance of an object detection model.