Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

IC15的训练结果计算hmean有点差 #20

Open
TuKJet opened this issue Apr 8, 2020 · 0 comments
Open

IC15的训练结果计算hmean有点差 #20

TuKJet opened this issue Apr 8, 2020 · 0 comments

Comments

@TuKJet
Copy link

TuKJet commented Apr 8, 2020

训练集只用IC15,默认代码里的300000次step;
loss普遍能下降几个数量级,但是结果不太好,使用IC15的测试集,计算hmean结果为:
{'precision': 0.023610427939006393, 'recall': 0.023110255175734232, 'hmean': 0.023357664233576644, 'AP': 0}

可以使用其他repo里的hmean计算脚本,注意修改bbox的数据代码
在./nets/utils.py中的unmold_detections函数中。将bbox的代码改为:

’‘’
points[:,[0,1]] = points[:,[1,0]]
rect = cv2.minAreaRect(points)
# (x1, y1, x2, y2, x3, y3, x4, y4)
boundbox = cv2.boxPoints(rect)
bound_boxes.append([boundbox[1], boundbox[2], boundbox[3], boundbox[0]])
full_masks.append(full_mask)
full_masks = np.stack(full_masks, axis=-1)
if full_masks else np.empty(original_image_shape[:2] + (0,))
bound_boxes = np.array(bound_boxes)
if bound_boxes else np.empty((0,8))
‘’‘

运行推理的时候记得删除gts
具体的计算脚本可以找其他repo中的cal_recall_precison_f1

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant