Parameter problem
Closed this issue · 0 comments
chenxyyy commented
Hello! @ghimiredhikura Thank you for your great work! I would like to ask two questions about the Parameter.
First:
utils/utils.py
def non_max_suppression_rotated_bbox(prediction, conf_thres=0.95, nms_thres=0.4):
"""
Removes detections with lower object confidence score than 'conf_thres' and performs
Non-Maximum Suppression to further filter detections.
Returns detections with shape:
(x, y, w, l, im, re, object_conf, class_score, class_pred)
"""
……
You said that the prediction is N × 9,but when I debug the eval_mAP.py
, I found that the prediction is N×10. I don‘t know why?
Second:
def rotated_bbox_iou_polygon(box1, box2):
box1 = to_cpu(box1).numpy()
box2 = to_cpu(box2).numpy()
x,y,w,l,im,re = box1
angle = np.arctan2(im, re)
bbox1 = np.array(bev_utils.get_corners(x, y, w, l, angle)).reshape(-1,4,2)
bbox1 = convert_format(bbox1)
Could you tell me the specific meaning of im and re? Is math.sin(float(yaw))
and math.cos(float(yaw))
?