need help in inference

[Liduye]

how can crop the detected image which is given to the model by using the code
!python3 main.py --mode=Image --Output=""--anchors_path='model_data/yolo_anchors.txt'--classes_path='model_data/cervix_classes.txt'\ --model logs/.h5
the code above gives an out put with bounding box and IoU score but I need to have croped image