1
$\begingroup$

I want to use “Fingerprint matching using multi-dimensional ANN” by Rajesh Kumar and B.R. Deva Vikram [content link] for fingerprint identification. But I have a serious problem understanding what is a Multidimensional Artificial Neural Network.

I searched for keywords MDANN and Multidimensional ANN but I can't find anything. I also can't be sure what Kumar meant about a MDANN in his article. So far, the closest thing I can relate with Kumar's MDANN is Convolution Nets but I don't believe this is what he meant. Can anybody help me about understanding this type of ANN?

$\endgroup$
1
$\begingroup$

Based on my skim of the paper, I suspect that this is just a standard neural network (in which the entire image is fed into the neural network as an input, one input line per pixel in the image), and they have invented a fancy-sounding name for a standard use of neural networks. You could of course contact the authors to ask them this question, but that's what it looks like to me.

To replicate their work, I suggest you read Section 5.1 very carefully and use the descriptions there to replicate their work.

$\endgroup$
  • $\begingroup$ What i couldn't get from their description was ; If you look at the multi-layered picture (fig 1) , a pixel on a column is connected to the corresponding neuron column's top and bottom neurons. I assume he just omitted drawing other connections for visuality. What i suspect is , is every single pixel is connected to every single neuron or is it like 1st column's every pixel is connected to every 1st column neuron. The second thing was the author saying "15x15 image and 20 neurons" yet Fig1 has 25 neurons and 5x5 image. So what kind of article writing is this ? Is it me or is it their writing? $\endgroup$ – Mert Çelikok Dec 18 '13 at 23:18
  • $\begingroup$ @MertÇelikok, yeah. I suspect the figure just omits other connections to avoid making Fig.1 too crowded. The paper says, "Fig. 1 shows the network structure of MDANN but the connections between the input and the hidden layer are not shown completely for the purpose of clarity". I find the paper ambiguous/unclear on the connectivity of the network. Your best bet might be to contact the author, or experiment some on your own. As far as the size of the image, good catch, you are right, I missed that. It's not just you; I find the writing unclear, too. $\endgroup$ – D.W. Dec 19 '13 at 0:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.