How to create training data set for signature verification?

2 次查看(过去 30 天)
I have 5 genuine and 5 forged signatures of each of 12 people( total 60 genuine and 60 forged). I've extracted their features like moments and aspect ratios (7 features of each signature). So, how do I create a training data set? Create a column for each signature? Do I need to create 12 training sets for 12 people? How do I arrange those features I've extracted to train the network ? (P.S : I've done some image processing projects but I'm a bit new in neural networks.)
Please help. Thanks.

采纳的回答

Greg Heath
Greg Heath 2018-2-13
[ 7 120 ] = size(input)
[ 12 120 ] = size(target)
target columns are columns of eye(12)
Hope this helps
Thank you forformally accepting my answer
Greg
  1 个评论
sushant nepal
sushant nepal 2018-2-26
I don't actually want to recognize signatures, i just want to verify them as in "legit or not", doesn't matter who it belongs to ... so does it really require a target data of 12 X 120 ?

请先登录,再进行评论。

更多回答(0 个)

类别

Help CenterFile Exchange 中查找有关 Sequence and Numeric Feature Data Workflows 的更多信息

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by