To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
A Novel Complex Feedback Independent Component Analysis Algorithm and its Application to Fingerprints Image Separation |
|
Author |
V. Singh, C. M. Markan, P. K. Kalra, V. G. Das |
| Citation |
Vol. 25 No. 6 pp. 81-85 |
|
Abstract |
Independent Component Analysis (ICA) is now a day become a more stable and sophisticated statistical method to analyze any multivariate data. There are various ICA algorithms already available to find the independent components as well as to separate the sources from the mixtures. Reviews show that researchers have focused more on separation of simulated data rather than the real life data. Most of the available algorithms are not able to completely solve the separation problem of the real life mixtures of images and audios. We have proposed a novel algorithm known as complex feedback ICA algorithm (complex-FEBICA), which is a gradient based algorithm with feedback architecture. It is well known that in the complex domain, rotational invariance can be found; complex-FEBICA is highly applicable to the real life mixtures. We have applied our algorithm to different fingerprint mixtures either created artificially or real life mixtures and have demonstrated the successful separation of fingerprints from mixtures. We are also able to separate m fingerprints out of less than m unknown mixtures. |
|
Keywords |
Independent component analysis (ICA), rotational invariance, complex domain, FEBICA, fingerprints. |
|
URL |
