Since the inception of biometrics, techno-crafts have been hard at work to make this technology securer and more efficient. Owing to the collaborative efforts, consistent development and regular improvements, biometrics came out to be what it is today. There were a number of challenges in giving equipment ability to capture, process and match biometric identifiers, which varied considerably from one individual to another. As biometrics came along, fingerprint recognition systems specifically saw widespread adoption and growth. They enjoyed attention from investors, tech experts as well as consumers. Being easy to deploy and use, fingerprinting soon became the most used method for personal identification in government organizations as well as private setups. Widespread deployments induced peer pressure among organizations, which resulted in more deployments. Use of fingerprint by governments for nation ID cards and civil identification applications gave fingerprint biometrics the seal of accreditation, which grew confidence of private and institutional setups like banks and financial institutions as well.
Historically, fingerprinting has been used mostly in crime scene investigations and forensics. Before the wave of information technology and digitization hit the world; fingerprint enrollment, storage and match were processed manually. Fingerprints were taken with ink and roll method on a paper sheet, and stored that way. This process was prone to errors and resulted in poor print quality. Current applications make use of live scan on modern fingerprint recognition devices. Live scan fingerprinting is the process of capturing fingerprints with electronic means, and without the use of ink and paper. Live scan is commonly used by law enforcement agencies and private facilities for identification, background check, criminal booking, etc. Live scan is quicker than ink and paper based scan, and results can be obtained much faster than tradition methods.
What is the result of a fingerprint live scan?
Result of live scan is a fingerprint image in digital form residing in device memory. This image carries all fine details of an individual’s friction ridges that can be used by the systems to uniquely identify or authenticate the individual. However, before this image could be leveraged for the purpose, it has to go through a couple of important steps: enhancement and compression. Image enhancement algorithms process the image and enhance quality of available details so that the recognition system can make the best use of it. Compression is required as captured image is usually of size that would not be feasible to exchange between different systems via the internet or other networks.
This exchange can take considerable time specially when network is slow or systems are overloaded with several requests. Despite the compression, there may be disagreement among devices that does not follow a common standard to exchange data. Different vendors may employ different technique to capture fingerprint information, the method of capture and the information to be exchanged. Fixing these issues required a common standard and this is where ISO/IEC 19794-4 came to rescue. It constitutes common standards for device manufacturers to follow, which supports interoperability and data interchange among biometric applications and systems.
What is image compression in general?
Image compression, as the name suggests, is a method that considerably reduces size of the image with or without sacrificing details. Loss of details can take place in a lossy compression formats like JPEG, while formats like PNG, BMP can compress images without losing details, hence, known as lossless compression formats. Loss of details depends on the compression ratio used in compressing the image in a lossy compression format like JPEG. Higher compression ratios can considerably deteriorate the image quality. (See the example image above).
Images captured by digital cameras, scanners and many other imaging devices are bigger in their original format. JPEG image compression algorithm compress these images to reduce size using low compression ratio to not to lose much details. Imaging devices like cameras, scanners, etc. keep compression ratios lower to keep details intact, on the other hand, higher compression ratio is used when an image is captured for the sole purpose of exchanging it over the internet, e.g. capturing an image for setting it as a profile picture on a chatting app. High compression ratio reduces image size, making it light on transmission channels, like the internet.
Why fingerprint image compression is necessary?
For the same reason a normal image compression is necessary. A fingerprint image captured by a recognition system is produced with a much larger size, making it heavy on resources like data storage and transmission channels. ISO/IEC 19794-4 encourages vendors of fingerprint recognition systems to initially capture a high quality image before compressing it, despite the difference in intermediate outputs produced by different approaches.
Fingerprint images can make use of a compression method which can compress them without losing important details. JPEG2000 is a popular image compression standard which can produce considerably smaller file size without losing major details. It is widely used for imaging devices like camera, scanners, etc. JPEG2000 can produce high quality images using low compression ratio. A 1:5 compression ratio produce higher quality images than 1:20, however, with high quality images, size increases. PNG (Portable Network Graphics is another graphic file format that support lossless compression. WSQ (Wavelet Scalar Quantization) is one of the popular compression algorithms for grey scale images. This algorithm was developed by the FBI, the National Institute of Standards and Technology (NIST) and the Los Alamos National Laboratory. It has become a standard for storing and exchanging fingerprint images over time. WSQ file format is used by most American law enforcement agencies.
Historically, fingerprints have been captured by ink and roll method, in which the finger is painted with ink and rolled on a paper sheet to capture fingerprints. Law enforcement agencies like FBI has been using this method since 1924, and now it has a huge amount of inked impressions on paper cards occupying an acre of filing cabinets. Conversion of paper card fingerprint image into digital one with high quality will cause image to occupy around 10MB of space. With millions of papers cards in FBI’s possession, digitization of all records will require several terabytes. These digital images will turn unusable due to their size. To render them light on resources like storage and transmission time, they need to be compressed. The same is the primary reason of compressing fingerprint images captured by live scan fingerprinting.