Proc. SPIE. 6982, Mobile Multimedia/Image Processing, Security, and Applications 2008
KEYWORDS: Digital signal processing, Biometrics, Independent component analysis, Visual process modeling, Data modeling, Neural networks, Algorithm development, Performance modeling, Systems modeling, Neurons
Among all existing biometric techniques, fingerprint-based identification is the oldest method, which has been
successfully used in numerous applications. Fingerprint-based identification is the most recognized tool in biometrics
because of its reliability and accuracy. Fingerprint identification is done by matching questioned and known friction skin
ridge impressions from fingers, palms, and toes to determine if the impressions are from the same finger (or palm, toe,
etc.). There are many fingerprint matching algorithms which automate and facilitate the job of fingerprint matching, but
for any of these algorithms matching can be difficult if the fingerprints are overlapped or mixed. In this paper, we have
proposed a new algorithm for separating overlapped or mixed fingerprints so that the performance of the matching
algorithms will improve when they are fed with these inputs. Independent Component Analysis (ICA) has been used as a
tool to separate the overlapped or mixed fingerprints.
Facial expressions are undoubtedly the most effective nonverbal communication. The face has always been the equation
of a person's identity. The face draws the demarcation line between identity and extinction. Each line on the face adds an
attribute to the identity. These lines become prominent when we experience an emotion and these lines do not change
completely with age. In this paper we have proposed a new technique for face recognition which focuses on the facial
expressions of the subject to identify his face. This is a grey area on which not much light has been thrown earlier.
According to earlier researches it is difficult to alter the natural expression. So our technique will be beneficial for
identifying occluded or intentionally disguised faces. The test results of the experiments conducted prove that this
technique will give a new direction in the field of face recognition. This technique will provide a strong base to the area
of face recognition and will be used as the core method for critical defense security related issues.