The AFIS features a new pre-processing technology on the market, dealing especially effective with low quality fingerprints that exceed established accuracy benchmarks. Due to these technical features, the accuracy for low and high quality fingerprints is exceeding industry standards and generating a solid Return on Investment for the end-user by achieving a industry leading performance.
Delfi can be scaled vertically & horizontally, can be extended in the future during its operation and has all the flexibility, robustness, salability, efficiency and accuracy a demanding authority can ask for nowadays from the industry. Delfi is fully integrated into our live enrollment system “BioCap” for de-duplication checks at time of enrolment, our Visa Managment Solution “Visalon” , our form based data capture solution “FormCap”, our Border Management solution “SBMS”, and our Central Register Systems CRS for large scale nationwide systems.
Key Features
Powerful Architecture
• SOA Multi-tier Architecture
• Truly scalable & distributable
• Advanced DB Replication
• Robust J2EE server side
• Web-Service (XML) Communication
• Hardware & Platform Independent
• Complex Workflow Orchestration
• Data-replication & Synchronization
• High Availability & FailOver cluster support with Load Balancing
• Backup Strategy for Disaster Recovery support
• DB Encryption support
• Support for Secure Communication & PKI infrastructure with AES 256 default support for highest security
High Flexibility & Return on Investment
• The Delfi AFIS can match effectively flat, rolled & latent fingerprints
• Due the support of rolled & latent fingerprints, Delfi has criminal capabilities and can be extended to a Criminal AFIS in the future.
Better Overall Quality & Accuracy
It is designed to deal effectively with low quality fingerprints including:
• Shift distortion
• Rotation distortion
• Elastic distortion
• Latent prints
• Scratches
• Low Contrast
• Smudges
• Problematic Skin Types
• Wet & Dry Fingers
• Low minutiae counts
• Partial Fingerprints
• Etc.
The better the fingerprint quality the higher the accuracy of matching this fingerprint in identification andd / or de-duplication scenarios - kindly refer to our QFinger flyer for more infromation.
Fingerprint pre-processing for higher accuracy
The state-of-the-art pre-/ post-processor for fingerprint images developed with medical image analysis background, produces enhanced images for higher quality enrollments and a thorough quality assurance process and delivers greater accuracy to every
AFIS system.
• Independent benchmarking by NPL in the UK and NIST has shown that the accuracy performance of the AFIS systems using the Preprocessor has improved significantly.
• The Preprocessor deals with the following fingerprint problems; shift distortion, rotation distortion, elastic distortion, scratches, low contrast, smudges, problem skin types, low minutiae counts and partial prints, producing enhanced images that can be matched with greater accuracy than before.
• The Preprocessor aligns each fingerprint as part of the enrollment process so that stored images are all pre-aligned, which enables faster matching speeds.
• The Preprocessor creates an idealized model of the fingerprint image, with the unique features preserved. This enables a significant improvement in accuracy, even when used with third party AFIS solutions, due to the better quality of the fingerprints to be matched
Official tests conducted by NIST show clearly the improvement of low quality fingerprints (NFIQ 4-5) to high quality fingerprints (NFIQ 1-2) in 90% of all cases, in a reliable fashion. This advanced pre-processing technique makes the Delfi AFIS performing outstandingly good in difficult environments and enrolment situations.
Proven World-Class Accuracy
The Delfi matching engine showed in all tests done by independent national & international organizations an outstanding performance within the top providers of the world.
NIST IR 7221 Test
Results in the table below demonstrate that the preprocessing and modeling technique has not only achieved the best average matching accuracy (AVG TAR), but more importantly, has also ensured high matching performance on low quality data (MIN TAR) compared to all leading AFIS providers.
NIST PFT Testing
- The NIST PFT test or NIST SDK vendor test as it was called in the past delivers another benchmark for the industry over another set of fingerprint databases for adequate performance evaluation.
- In the NIST PFT Test at 0.1% FAR Avalon resides amongst the Top 3 with Cogent & NEC.
NIST PFTII Testing
- Currently Avalon Biometrics is the only company that participated in the NIST PFTII test and has its results published.
For details see http://www.nist.gov/itl/iad/ig/pftii_results.cfm
