search

szagoruyko / cvpr15deepcompare

Code and models for "Learning to Compare Image Patches via Convolutional Neural Networks"
http://imagine.enpc.fr/~zagoruys/deepcompare.html
Other
468 stars 169 forks source link

how to test a pair of image #36

Open sunbin1205 opened 6 years ago

sunbin1205 commented 6 years ago

Thank you very much for sharing! If I only want to test the similarity of two images in a pair of images, the return value I want is only the similarity probability of this set of images, can your code get such output? Which variable is the percentage of similarity that holds each set of test data that I didn't find in the code. Please specify that for me,this will be very helpful to my study! Thank you very much!

szagoruyko commented 6 years ago

Hi @sunbin1205, see https://github.com/szagoruyko/cvpr15deepcompare/blob/master/pytorch/eval.py

JiaqianLi commented 5 years ago

Hello, I have the same problem with @sunbin1205 and I want to know how to run the code aboout eval with two input images! Looking forward for your reply, thank you~!

szagoruyko commented 5 years ago

@JiaqianLi check these lines https://github.com/szagoruyko/cvpr15deepcompare/blob/master/pytorch/eval.py#L220-L230

JiaqianLi commented 5 years ago

Thanks a lot. I run the code and have another problem. I want to know how to organize the test data. whether I create a new folder named liberty and then put two test images can work? And the test_matches file will automatically produce?