Skip to content

jatinjindalj/kaggle-denoising-dirty-documents

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Kaggle Denoising Dirty Documents

My solution in this Kaggle competition "Denoising Dirty Documents", 6th place.

Model

Deep Convolutional Neural Network

Layer Type #Input Channel #Ouput Channel FilterSize Padding
Input - 1 - -
Convolution(1) 1 96 3 1
Leaky ReLU 96 96 - -
Convolution(2) 96 96 1 0
Leaky ReLU 96 96 - -
Convolution(3) 96 96 3 1
Leaky ReLU 96 96 - -
Convolution(4) 96 96 1 0
Leaky ReLU 96 96 - -
Convolution(5) 96 96 3 1
Leaky ReLU 96 96 - -
Convolution(6) 96 1 3 1
Output 1 - - -

Sample

Test Input
sample_test_input

Test Output
sample_test_output

About

6th place solution in "Denoising Dirty Documents"

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Lua 100.0%