A new energy function is proposed for forming self-adapting ordered representations of input samples in a multilayer perceptron. Simulation results on unconstrained handwritten digit recognition give a better invariance extraction for this model than for several other models.
References
-
-
1)
-
G. Wallis
.
Using spatio-temporal correlations to learn invariant object recognition.
Neural Netw.
,
9 ,
1513 -
1519
-
2)
-
B. Olshausen ,
D. Field
.
Emergence of simple-cell receptive field properties by learning a sparsecode fornatural images.
Nature
,
607 -
609
-
3)
-
J. Stone ,
A. Bray
.
A learning rule for extracting spatio-temporal invariances.
Netw.
,
429 -
436
-
4)
-
G. Wallis ,
R. Baddeley
.
Optimal unsupervised learning in invariant object recognition.
Neural Comput.
,
4 ,
883 -
894
-
5)
-
P. Földiák
.
Learning invariance from transformation sequences.
Neural Comput.
,
194 -
200
http://iet.metastore.ingenta.com/content/journals/10.1049/el_19980161
Related content
content/journals/10.1049/el_19980161
pub_keyword,iet_inspecKeyword,pub_concept
6
6