the effect between cross-brightness inputs

Question

the effect between cross-brightness inputs

CallMeFrozenBanana opened this issue 5 years ago · 2 comments

CallMeFrozenBanana commented 5 years ago

Hi，I used the model trained on synthetic data to test on the synthetic data. When the light levels of the input images are different, the model's prediction effect is significantly worse than the ordinary case. How do I get the effect described in the paper?

Answer 1 · 2019-12-10T23:04:29.000Z

We also found that the brightness is an important factor. If the two images have a huge diff in the brightness, you may expect a worse performance like you said. A trick is to normalize the two images to have an equal mean of intensity, for example.

…

On Tue, Dec 10, 2019, 5:38 AM frozenbananana ***@***.***> wrote: Hi，I used the model trained on synthetic data to test on the synthetic data. When the light levels of the input images are different, the model's prediction effect is significantly worse than the ordinary case. How do I get the effect described in the paper? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#20?email_source=notifications&email_token=ABKAZDJ25AIEJIYCV7WMYTTQX5WRHA5CNFSM4JY4IL6KYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4H7NHTBA>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABKAZDL4QJISY7BP2J3GPGDQX5WRHANCNFSM4JY4IL6A> .

Answer 2 · 2019-12-17T09:49:05.000Z

We also found that the brightness is an important factor. If the two images have a huge diff in the brightness, you may expect a worse performance like you said. A trick is to normalize the two images to have an equal mean of intensity, for example.
…
On Tue, Dec 10, 2019, 5:38 AM frozenbananana @.***> wrote: Hi，I used the model trained on synthetic data to test on the synthetic data. When the light levels of the input images are different, the model's prediction effect is significantly worse than the ordinary case. How do I get the effect described in the paper? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#20?email_source=notifications&email_token=ABKAZDJ25AIEJIYCV7WMYTTQX5WRHA5CNFSM4JY4IL6KYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4H7NHTBA>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABKAZDL4QJISY7BP2J3GPGDQX5WRHANCNFSM4JY4IL6A .

Get it！thanks~