June 20, 2019

NEC successfully carried out the world's first proof of concept of fingerprint identification of newborns in the Republic of Kenya. We interviewed a principal researcher, Yoshinori Koda to learn about his motivations and results including on the significance of fingerprint identification of newborns and how this technology contributes to society.

Providing required public services to children around the world

― Why is fingerprint identification of newborns important?

Thus far, I have been involved in biometric authentication technologies including fingerprint identification for more than 20 years. Among those technologies, National-ID systems featuring a biometrics identification function are a major type of system used in many countries. For example, in India, biometrics information is used together with a unique ID number and enables more than one billion individuals to establish and prove their identity. I am delighted that my involvement in biometrics technologies is contributing to building a society that provides social services to every citizen in a fair manner.
The Sustainable Development Goals (SDGs) of the 2030 Agenda for Sustainable Development, adopted at the UN summit aims to eradicate poverty in all its forms and dimensions. Target 16.9 of Goal 16 requires that states should, “by 2030, provide legal identity for all, including birth registration.” In fact, the UN estimates that as many as 1 billion people struggle to prove who they are (*1). This means that there are still many people around the world who do not possess a legal identity. Without a legal identity which originates from a birth registration, a person loses the opportunity to receive various public services such as education, medical treatment, and welfare which are the fundamental rights of a citizen of that country. For example, if a person cannot prove that they are a citizen of that country, they cannot acquire a passport or travel overseas. Moreover, this can also lead to the situation where they cannot be granted employment opportunities which can become a factor in preventing them from escaping poverty.
I believe that biometric authentication technologies can contribute significantly to improving this situation. From a systematization perspective, there are regions of the world in which the management of citizenship by governments is still lacking. Furthermore, ID cards have low credibility due to rampant forgeries in some cases. Because there are no concerns about loss, and they are highly reliable, biometric authentication technologies are extremely effective in regions that hope to solve complex issues using cutting-edge technologies. In addition to India as a major use case, NEC has also delivered national identity systems to countries such as the Republic of South Africa.
However, the fact of the matter is that many national identity systems do not include newborn children, infants, or young children within their target age range (*2). This is due to various issues such as the technical difficulty of covering the biometrics information taken from children, but as far as I am concerned, I continued to believe that solving the technical issues would somehow overcome these barriers. The reason is that performing a birth registration immediately after birth is extremely important for securing a legal identity to prove one's identity in the future. Moreover, a baby cannot say their own name or present an ID. However, you can physically prove the identity of a baby with biometric authentication. We believed that the strongest feature of biometric authentication should be used to the maximum extent to protect babies from environmental deprivation and ill will.
It is said that 5.6 million children under the age of five lose their lives around the world. In addition, 2.8 million of those children die within one month of their birth (*3). One of the underlying issues is the problem of preventive inoculations by vaccination not being properly carried out. Newborn children receive many vaccines up to 14 weeks after birth, and it is recommended that they be inoculated at the proper time . Each country is diligently working to protect the lives of children. However, the planning and historical management of inoculations at the local level is still insufficient. In some cases, they are unable to deliver the vaccines to children because they do not know who has been vaccinated, when they were vaccinated, or what vaccines they received or still need. In addition, even if the vaccines are delivered, sometimes the families are unable to afford the bus fare to go to the hospital. In such cases, United Nations organizations and NGOs will go to the children, but they do not know who those children are. In the worse cases, their parents may have died, and the agencies are unable to verify the identities of children who need to go to a hospital.
At the request of some organizations of the United Nations, I myself have provided technical assistance for identity verification in activities to support developing countries. I received a request from a researcher at a government agency in a certain country who wanted to ensure that the appropriate medical treatment was provided to children who may have been infected with HIV through their mothers and verified the feasibility of identity verification using biometric authentication. In such environments, I felt the gap between the smiles of the children and the reality they are living in. I witnessed such conditions with my own eyes and ears. Such experiences made me think that I had to implement fingerprint identification of newborn children as an effective way to overcome such international issues.
The idea was to build an environment that ensures birth registrations to provide all children with a legal identity regardless of the country or region where they were born by using fingerprint identification technologies that can precisely prove who a child is just by having them extend a finger. The next step was to build an environment that allows the children to reliably and properly receive the public healthcare, educational opportunities, and social insurance needed during their growth process which they should receive as citizens. I believe that the research needs to be properly developed in order to realize an equitable society as well.

Overcoming difficulties to achieve the world's first fingerprint identification of newborn children

― Why was fingerprint identification of newborn children not realized until now?

Fingerprints have uniqueness and eternity as its features. Therefore, the fingerprints of newborn children can be imagined as the miniature version of their adult fingerprints when considered as a pattern. However, the established research on fingerprints for newborn children was limited, and the verification data did not exist. In order to implement fingerprint identification of newborn children, it was necessary to first understand what exactly the fingerprints of newborn children are.
So, the first thing that I did was to contact professors in the medical and nursing schools of various universities to explain my ideas and ask them to teach me about newborn children. Because this was still the era of MDGs (Millennium Development Goals) before the adoption of SDGs, such international issues were not as well known, and people were confused about the motivation behind this research. Now, I can look back fondly on those times.
Next, I needed to collect fingerprint data for newborn children. However, I knew from my preliminary investigation that the roughly 500 ppi (pixel per inch) resolution of data collection equipment generally used for fingerprint identification of adults was unsuited to imaging the fingerprints of children. Additionally, the high-resolution imaging equipment for properly scanning the fingerprints of newborn children did not exist. Because the equipment did not exist, I would have to create it. Through a repeated process of trial and error to evaluate various sensors and imaging methods, I built a prototype to acquire the data. I would like to take this opportunity to express my gratitude to the senior colleagues who provided guidance in developing that prototype.
After the prototype was completed, I was blessed with an opportunity to collect a somewhat small set of data. As a result of using the data to verify the fingerprint characteristics of newborn children, I realized that the surface of the skin was not simply scaled down in size but possessed various characteristics. I also realized that these characteristics increased the difficulty of fingerprint identification.
An easy way to understand it is that the extremely small size of their fingers is an issue. When trying to scan the fingerprints of a newborn child with sufficient accuracy, you must be able to clearly scan the less than 20 µm lines which make up the same fingerprints of an adult within an area that is approximately 1 cm square. Therefore, we used a high-resolution CMOS sensor, but that alone was not enough to scan the very sweaty fingerprints of newborn children. As a result, we use a method which applies a special type of glass patented by NEC to the surface of the sensor to clearly capture the fingerprints. Based on the results of previous testing trials, a resolution of 2400 ppi was adopted as the dedicated configuration for newborn children, and the scanning was successful.

Of course, it goes without saying that babies are unable to obediently place their fingers on a scanner in front of them. The on-site operator must bring the handheld scanner close to the baby's finger and place it on the scanning surface. However, babies will reflexively try to clench their hands and others may get scared and cry, so there was always a risk that the fingerprints would come out blurry.
Furthermore, babies have soft skin and a large amount of moisture inside their bodies. Therefore, just placing their finger on the scanning surface, pressing down lightly, and moving it leads to multiple wrinkles on the surface of their finger, and the fingerprint is scanned with a deformed shape. This "elastic deformation" phenomenon was one of the major hurdles which made it difficult to implement fingerprint identification of newborn children. For that reason, we attempted to solve the problem by increasing the scanning frame rate to 7 fps (frames per second). Previous fingerprint identification tool was performed at 2 fps, so you can see that this is a significant increase in speed. In addition, we also approached the prevention of elastic deformation from a hardware perspective as well. We designed the scanner to prevent fingerprint smearing even if the finger is pressed down with excessive force by attaching walls with the appropriate height on either side of the scanning surface. In addition, these walls restrict the horizontal width and depth of the imaging surface and make it difficult for the finger to move when placed on the surface. We implemented this design based on the idea that placing near infrared LEDs on both sides of the imaging surface in an effective manner would help deal with the problem of elastic deformation. We were able to achieve fingerprint imaging of newborn children with a consistent level of quality by integrating the optimal hardware and software including clever ideas such as these.

Enabling authentication two hours after birth to allow the mother and child to leave the hospital within six hours after giving birth

― How soon can a child be registered and authenticated after being born?

The basic configuration of this scanner was implemented to a certain degree during the prototyping stage in 2016, but at that time, we still did not know to what extent it would be able to perform. Nevertheless, we successfully scanned the fingerprint patterns of newborn children six hours after birth during the testing trials.
However, by implementing features which are specific to newborn children with these improvements, we demonstrated that we can scan even newborn children as soon as two hours after birth. That might seem like just a reduction of four hours, but the fact that we were able to achieve that reduction through testing trials is extremely important. The reason why is that we noticed during testing trials in Kenya that depending on the region, many mothers and infants checked out of the hospital less than six hours after giving birth. We were told that the reasons for this behavior included customs in which the mother and newborn child would relax at home after the birth, the support of relatives, and the extremely strong local ties. Moreover, the hospital environment is different from Japan. For example, there are no air conditioners in the hospital rooms, and sometimes even the fans stop working. Depending on the room, the windows may not close or lack screens, so there is also a risk of malaria infection. If the environment is essentially the same as being at home, many mothers decide to go home and check out of the hospital within about six hours. I was surprised to see mothers walking in front of me carrying their belongings immediately after giving birth.
Therefore, being able to register the fingerprints of newborn children without fail within six hours after their birth carries great significance. It enables the verification of the child's identity when checking out to prevent misidentification and to create birth certification data which properly certifies that the child was born in that hospital. If you can certify what hospital a child was born in and the nationality of their parents, that information can be tied to a birth registration as a citizen of that country to provide a legal identity to all babies. Regrettably, nations cannot just recognize any child that is brought to a government office as a citizen. Some countries in Africa are facing a refugee problem, and the situation is extremely serious.

Developing the unprecedented field of fingerprints for newborn children

― What are your future goals?

The fingerprinting of babies is a field that has never really been researched at all. There have been several reports stating that typical fingerprint identification can be applied starting from 12 months after a child is born. However, there was little prior research on the scanning of fingerprints within six months after birth, so it was unknown territory. At the current moment (this interview was conducted on May 23, 2019), it has only been 14 weeks since the start of this research. How do children's fingerprints change as they grow? How are they different from the fingerprints of adults? I believe that this research holds great significance in terms of figuring out such issues through the use of electromagnetic recordings. Currently, we are only able to recognize fingerprint samples during the fragmented period of newborn children, so I would like to continue to be able to properly recognize fingerprints at the time of birth and then at age one, age five, etc. In addition, I would like to verify that the fingerprint patterns do not change from the time immediately after birth using electromagnetically recorded data. Creating such data will absolutely require continuous research, so I would like to focus on that going forward.
Of course, the ultimate objective is to be able to properly verify the identities of newborn children and infants in a persistent manner at the time when it is required. The current research focused on fingerprints, but I also expect to apply face recognition and other modality of biometric authentication technologies. I hope to continue working hard to apply biometric authentication to implement birth registrations so that every child will be provided with a legal personality and legal identity.

Go back to Featured Technologies