MATLAB: How to train labels for image blocks using neural network

Deep Learning Toolboxlabelingneural network

my input is a matrix each row filled with image block values and another matrix (column vector)having labels for each block. I have to train each block with the label using neural network.

how to implement this? I am new to neural network, please any one help me.

Best Answer

Form 2 matrices

input columns are columnized matrices (which may be the result of using feature extraction on the original matrices)

target matrices contain the label columns.

 help fitnet

However, if this is a classification/pattern-recognition problem, the target columns should be columns of the unit matrix.

help patternnet

Hope this helps.

Thank you for formally accepting my answer

Greg

Related Solutions

MATLAB: How to train a feedfordward neural network with error weights

 Use the command window
 help mse
 doc mse

Hope this helps.

Thank you for formally accepting my answer

Greg

MATLAB: How to use non-integer(decimal) values as target values in neural network

For N pairs of I-dimensional "I"nputs and corresponding O-dimensional "O"utput targets,

 size(input)  = [ I N ]
 size(target) = [ O N ]

For regression/curve-fitting see the FITNET documentation with sample code:

 help fitnet
and
 doc fitnet

For classification/pattern-recognition see the PATTERNNET documentation:

 help patternnet
and
 doc patternnet

For more examples you can search the NEWSREADER and ANSWERS using

 greg fitnet
or
 greg patternnet

Hope this helps.

Thank you for formally accepting my answer

Greg

Best Answer

Related Solutions

MATLAB: How to train a feedfordward neural network with error weights

MATLAB: How to use non-integer(decimal) values as target values in neural network

Related Question