search

kevinlin311tw / caffe-cvprw15

:heart::coffee: Deep Learning of Binary Hash Codes for Fast Image Retrieval (CVPRW15)
Other
526 stars 218 forks source link

After run demo.m? How to achieve Image Retrieval? #26

Closed lshignal closed 6 years ago

lshignal commented 7 years ago

I found that after running demo.m, I got binary_codes of images.

However, how can I achieve Image Retrieval after getting binary_codes? It seems that your code doesn't contain image retrieval part. @kevinlin311tw

kevinlin311tw commented 7 years ago

You can find image retrieval part as well as the evaluation in run_cifar10.m

2017-09-13 1:56 GMT-07:00 lshignal notifications@github.com:

I found that after running demo.m, I got binary_codes of images.

However, how can I achieve Image Retrieval after getting binary_codes? It seems that your code doesn't contain image retrieval part. @kevinlin311tw https://github.com/kevinlin311tw

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/kevinlin311tw/caffe-cvprw15/issues/26, or mute the thread https://github.com/notifications/unsubscribe-auth/AJgtMEXfwFVh5SGD_p2-qf_CadDegfreks5sh5jVgaJpZM4PVyS5 .

--

Best regards,

林可昀

Kevin Lin

brijesh099 commented 7 years ago

@kevinlin311tw I am new to this field and do not know much about caffe. I just tried to execute your caffe model with run_cifar10.m and got MAP of 0.89. I couldn't figure out how it is taking input(query) image and provides the image retrieval. I can only see that, it provides MAP and precision file based on training and test sets. Can you please provide steps on how to provide query image and retrieve the images from data set?

Thank you in advance for your time and help.

kevinlin311tw commented 6 years ago

@brijesh099 We take each image in the test set as the query. We retrieve similar images from the training set.

Step 1, Get the binary code of the query image. See: https://github.com/kevinlin311tw/caffe-cvprw15/blob/master/analysis/precision.m#L17

Step 2, Compute similarity between the query image and the images in the gallery. See: https://github.com/kevinlin311tw/caffe-cvprw15/blob/master/analysis/precision.m#L20

Step 3, Rank and find the nearest neighbor. See: https://github.com/kevinlin311tw/caffe-cvprw15/blob/master/analysis/precision.m#L30