Fusing with Context: a Bayesian Approach to Combining Descriptive Attributes

Fusing with Context: a Bayesian Approach to Combining Descriptive Attributes

The increasing demands and claims for the high and accurate intelligence, forensic and surveillance systems propelled the unimpeded face identification issues and problems to support the biometric verifications and research. During the last decade, exceptional progress has been made to solve and resolve the controlled and uncontrolled face verifications issues. But only low quality of improvement and advancement has been achieved in this regard.

Verification of the face is quite easy issue and problem than the face identification. As the face verification, only requires the discrete pairs of images for the matching purpose. Whereas, identification has been developed in to a problem in the situation when an unknown identity from the data and set of known identities.

In case of the identity problems, descriptive attitudes can be used and can take the formula of any information that can be used to identify the individual such as data, age, contextual data and the describable and tangible physical attributes with a set and description of the descriptive attributes of the individuals, the advancement and improvement in the traditional and conventional face identification system by intelligent weighting of the scores.

If such a system will be developed in which the someone can control the different and district attributes of the people than the system can detect the and trace the incorrect search results and findings and will be able to identify that how could the correct record can be build that will give the high matching scores. It is so natural that the all provided and searched attributes cannot be matched. This paper and research the Bayesian networks have been identified and searched and with the help of Bayesian networks the more accurate and descriptive outcome has been achieved.

In the framework of these Bayesian networks with the help of the description of the attributes the appropriate and the probability of the relevant searches can be achieved and accomplished.

The contributions of this research, will support and extent the scope of the biometric searches. The Bayesian method also incorporates and provides information that is beyond the scope of the biometrics. In this research the extent of accuracy of the Bayesian method has also been mention and determined. This research has also determined the success of the Bayesian method when unknown attributes has been searched.

This paper has addressed the following issues:

It has introduced the Bayesian formula.

The advantage of this method is that it has introduced and integrated the descriptive attributes and has also incorporated those attributes which are beyond the scope of the soft biometrics.

This process is also preferred because it also extra information that are even not required by the biometric data.

The method has also introduced a Noisy-OR formulation for modernized value assignments and more perfect and accurate weighting.

The accuracy of the Bayesian method has shown that the best possible results. It has estimated the right age estimation.

Clearly with respect to the future work to enhance and expand the describable visual attributes and it value is a work in progress.