About your prepared Davis stationary masks sample numbers

[unclebuff]

Hi, thanks for your sharing.
关于您提供的 prepared stationary masks中，我注意到STTN方法中将Davis训练集90个场景的sample都用于evaluation，而您只从中选择了50个场景制作了mask. 请问这是出于什么原因考虑呢？是否是处于效率考虑，增加的场景不会影响evaluation结果？

另外，我还注意到在对比实验中，您直接使用了其他对比方法在STTN工作中的evaluation结果，您是否使用相同的验证集进行了验证并得到了相同的结果呢？

谢谢您的回答

[ruiliu-ai]

Since the DAVIS dataset is updated, the current DAVIS dataset contains more examples than what we use.
But all the examples we used could be found in current DAVIS dataset. The examples we used could be found in https://drive.google.com/file/d/1wihArvScAFT9hs3KDGSoCEjsEjXpfEOL/view
You can filter them out from current DAVIS dataset.

[unclebuff]

Thanks for your reply. What about the second question?

[ruiliu-ai]

Since STTN doesn't release their evaluation masks, the evaluation results are close instead of exactly the same.

[unclebuff]

Thanks for your reply.