Skip to content

linjianz/tensorflow_homographynet

Repository files navigation

Homographynet

This is an implementation of the paper Deep Image Homography Estimation with tensorflow
网络参数下载地址 百度网盘

Dataset

ms-coco

dataset image numbers
train2014 82783
val2014 40504
test2014 40775

Program list

train_mycnn.py

Build the network according to the paper completely.
No need to pre-save the generated data, the program genetated the image pairs automatically.

test_mycnn.py

Test only one image a time, output the four-pair offsets predicted from HomographyNet.
Then use data_process.m to visulize the results.

train_net.py

Use existing cnn framework for training, something wrong v.

data_generation.py

If you like, just generate the training data.

Result

I test 200 images on test2014, Mean Corner Error = 12.6578 (image size is 320x240).
The original thesis is 9.2 (image size is 640x480). But I believe my result could be better.
good example:
1

my result Vs. the author's result on the same image

bad example:

About

A tensorflow implementation of Homographynet

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published