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

作者,您好,请问 LSM-HAWP wireframe detection model 仅仅是在ShanghaiTech数据集训练的而在所有实验数据集上公用的吗? #5

Open
micklexqg opened this issue May 12, 2022 · 3 comments

Comments

@micklexqg
Copy link

如题,如果LSM-HAWP wireframe detection model 只在ShanghaiTech数据集训练,不在其他数据集上训练,会不会影响在其他数据集上的实验效果?谢谢!

@ewrfcas
Copy link
Owner

ewrfcas commented May 12, 2022

我们只在ShanghaiTech上训练,在其他数据集上表现也比较良好。因为主要也是针对建筑方面的情况,自然场景直线比较少。另外可以借鉴SOLD2这种自监督方式来到其他数据集上finetune。
https://openaccess.thecvf.com/content/CVPR2021/papers/Pautrat_SOLD2_Self-Supervised_Occlusion-Aware_Line_Description_and_Detection_CVPR_2021_paper.pdf

@micklexqg
Copy link
Author

谢谢!有个困惑,为啥论文里的在 comprehensive Places2 (P2C) (随机挑选的10个类别带自然场景和man-made场景的数据)比在man-made Places2 (P2M)(10个类别只包含man-made场景的数据)上效果好(Table 1),man-made Places2 (P2M)数据更贴近建筑方面,效果应该更好才对

@ewrfcas
Copy link
Owner

ewrfcas commented May 13, 2022

这可能是因为P2M的数据结构较多,相比自然场景的P2C本身难度就更高的原因

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

2 participants