There are many elaborated algorithms for fingerprint identification, but none of them can guarantee that the results of identification are always 100 % accurate.
The research published in B.R.A.I.N. journal, shows an initial phase in a fingerprint image analysing process that comprises in the pre-handling or filtering. On the off chance that the outcome after this stride is not by a decent quality, the forthcoming ID procedure can come up short. A major difficulty can show up if there should arise an occurrence of unique mark recognizable proof if the pictures that ought to be distinguished from a fingerprint image database are noisy with various sort of commotion.
The decision about the best filtered pictures of an arrangement of 9 algorithms is made with a double technique of fuzzy and aggregation model. The researcher is proposing through this paper an arrangement of 9 filters with various oddity intended for handling the computerized pictures utilizing the accompanying strategies: quartiles, medians, average, thresholds and histogram equalization, connected everywhere throughout the picture or locally on small areas. At last the measurements uncover the arrangement and positioning of the best calculations.
The identification of individuals based on biometrics are made by different methods like: face recognition, palm recognition, voice recognition or handwriting recognition but one of the most usual approaches consists in the fingerprint based identification.
Some of the most important implementations of fingerprint algorithms are the IAFIS-Integrated Automated Fingerprint Identification System used by the FBI, which will be followed by the Next Generation Identification system developed by Lockheed Martin in partnership with Safran, and also commercial applications implemented in: mobile phone applications, different operating systems like: Apple-iOS, Android, access systems and airports.
Automatic identification based on fingerprints can be synthesized into five distinct phases: pre-processing, feature extraction, feature measurement, classification and matching. At the step of pre-processing the challenge is to find the best methods for enhancing images, taking into consideration the accuracy and the time for execution.
Designing the model included three noteworthy steps: characterizing the arrangement of channels modified for the arrangement of unique mark pictures, characterizing the equipment for the programmed choice of the best filter for every situation lastly the order in the wake of handling an arrangement of pictures. In the annex, the researchers uncovered to a limited extent the aftereffect of every channel and the assessment and picking the picture that will be utilized as a part of ensuing phases of examination inside the significance of robotising the whole procedure of ID taking into account fingerprints. As a continuation of the computerization process there is planned to assess the fuzzy and statistical parameters with more advanced strategies such as neural systems.
Read more here!