Hi,

When are you planning to release your model? Is it possible to get it privately?

@yuruntian where we can find the paper for this repository?

Hi @yuruntian! I'm trying to reproduce your results using tensorflow. I have some questions about training process of L2-Net.

1. The author state in the paper that 2 is the maximum L2 distance between two unit vectors. Is there any normalization for the intermediate features or the output descriptor before calculating the loss function? I'm trying to reproduce your results using tensorflow. However, the loss value sometimes would be NaN without any normalization for the intermediate features or the output descriptor.

2. Did you use any max pool layer on your L2-Net? The size of output descriptor is 128. However, the size of features maps after the final batch norm layer without any pool would be 8x8x128. How to transform these features map to 128 descriptors？

Thanks,
Guangyao

there are some statements needed to make sure:

1. while the weighting and bias are fixed at 1 and 0 respectfully,### the mean and var in it should be learned in the training process. Right?

2.the author state in the paper that the bn layer before DIF is not learned. Then it is the mean and var are calculated every time for different input?

I am really interesting in your L2-NET work, and want to reproduce the training process.
Is the training code available or can I get access to it?

Thanks a lot.

Paper writed this sentence "we use a Local Response Normalization layer (LRN) as the output layer to produce unit descriptors."

So, I want to know is it definitly get a unit feature vector from the LRN ?

Hi, your work is great and I was inspired, but I have a little problem when I read the paper,Why is gkk the smallest in the third term of the loss function?Shouldn't the inner product between the features of the matching block be the largest?

1.What is the Browndataset URL?
2.Since I have not been in touch with matlab to realize deep learning, please tell me the specific operation instruction of this project.
thanks.

Hi @yuruntian, I read your paper and found it very interseting.
I'm trying to reproduce your results using tensorflow.
specifically,I'm trying to take the model trained on HPatches(with augmentation) and test it on Brown dataset.
I ported the weights from matconvnet into tensorflow,and followed the exact architecture.
The tensorflow descriptor works quite well for feature matching tasks, so I'm guessing I plugged in the weights correctly.
I also followed the Brown evaluation method and report FPN @ recall=0.95
in this case, however, I'm getting quite different results:
20% FPN @ recall=0.95 on liberty(vs. 3.2% you reported in the paper)
So I guess I must be doing something bad in the evaluation code.
Can you share or point me to the evaluation code you were using?
Also, can you elaborate more on how you measured yourself on brown dataset (patch size, special tweaks you had to do, etc..)?

Thanks,
Guy

I find Hpatches datasets already,but Brown datasets is hard to find.Where I can download this datasets?
Thanks!