The selfie was introduced into the English language on September 13, 2002, and in 2013 it was chosen as the word of the year by the Oxford dictionaries, marking a significant journey for an action that was previously known as a self-portrait. Nowadays, an average of 93 million selfies are taken worldwide every day. And now, financial and governmental institutions are joining in, requesting a self-portrait to validate your identity by comparing and recording your facial biometric parameters.

There are differences between a self-portrait for biometric validation and a selfie for social media. Selfies for social media are images that express the moment they are captured: travels, events, celebrations, or any situation where we want to capture ourselves within the landscape. Additionally, the camera position, lighting source, the quality of light affecting our faces, and the cleanliness of the lens don't matter. This is why there are so many bad selfies on social media. I dare to think that the type of selfie a person posts on social media says a lot about their character.

Facial recognition for biometric analysis is an image or frame from a recorded video when the person is facing a camera. This recognition is carried out by comparing the distances and relative shapes of a person to an existing image in the file.

To ensure a successful image capture and avoid identity problems during the biometric analysis of our face, I want to give you some suggestions on how to take a self-portrait for facial recognition. Aspects to consider: the camera, lighting, the environment around us, and last but not least, ourselves.

The Camera

The cameras on our desktop devices, laptops, and smartphones have different advantages and disadvantages. If we are not aware of their capabilities and limitations, we may have many difficulties capturing a successful biometric verification selfie.

Webcams on laptops are the worst and require more care to take a good picture. The camera's positioning is important; if we are using a laptop, we should ensure that the camera is at the same level as our eyes. For this, we can place the laptop on some books to gain height and be at the same level. The same applies to external webcams: maintain eye level with the camera, and our head should not be above or below the camera. In the case of cell phones, we should use the front camera to monitor our position on the phone screen and apply the same principle: our chin should be level, not looking down or up, and adjust the camera by raising or lowering the arm to be well centered. And lastly, but extremely important: The camera lens. Before starting the capture process, we must clean the camera lens; this is of utmost importance. Phone cameras are the most prone to getting dirty since they are constantly and extensively handled. In the case of computers, we simply never clean them. Dirty lenses cause loss of sharpness in a process where the image must be clear and clean.

Lighting for Our Biometric Portrait

Many banks and institutions say "Well-lit face." This is a very broad and imprecise concept, but institutions are not obligated to provide a lighting course to all users. What they mean by this is that the light hitting your face should have certain characteristics and qualities regarding its direction and intensity.

The best direction for taking a biometric self-portrait with your webcam or cell phone is to have the camera directly facing your face on the same camera axis. The current example we all know is the famous influencer ring light that is very fashionable these days. If you don't have one, it will be sufficient to turn your face towards a window to receive even soft and ample lighting that draws your face well without shadows.

Backlit images are not recommended because even if the camera captures your face, it will appear dark and without detail, making biometric verification impossible.

Another big problem is blurry images, which occur due to low light intensity. It is important that the light intensity is strong enough for our face to appear sharp. A blurry image will not be accepted for use. If you don't have natural light and decide to do a procedure at night at home, you must consider several factors. Warm or yellow-tinted lights must be avoided at all costs. Frontal light is still the best option, but as white as possible and with good intensity.

Our Environment

This part refers to the other elements that make up the image where our face is. It is important that our environment be as clean and simple as possible. Complex or dark backgrounds should be avoided, or even worse, both together. It is also very important to avoid other people appearing in the image, even if they are out of focus or partially within it. Mirrors cannot appear in the image, especially if they are reflecting our back or very bright surfaces. No type of pet, whether partially or out of focus, should be included—the cat should not participate in the selfie!

Ourselves

Finally, the most important thing is ourselves. Personal accessories are all contraindicated: headbands, combs, ribbons, or any other object to adorn the hair. The hair cannot obscure any part of our face; it must be properly tied back to not cover the ears or forehead. We have to avoid wearing sunglasses or corrective glasses. Some systems allow successful identification of people wearing glasses, but it is advisable not to use them since the eyes are a very important part of successful and accurate biometric identification. Some people have very shiny or sweaty skin. In this case, we can wash our face before taking the selfie to remove reflections. Regarding makeup, excessive makeup is contraindicated, especially colorful eyeshadows, as they can deceive the system into making errors. Our face should occupy between 70% and 80% of the image's surface. The space occupied by the face in the image should be close enough to see the details but also far enough to see the complete silhouette down to the neck. Very importantly, when capturing the image, we must maintain a neutral expression. Nowadays, although facial recognition technology is much more advanced, there is still a long way to go. New laser surface scanning systems (LiDAR, Laser Imaging Detection and Ranging) will make everything simpler as they have better surface reading capabilities than a photographic camera.

For now, it is important to follow these small details, not only for identification "per se," but also for our own safety on the web. A better capture of our biometric data will provide greater personal security in the digital age.

‍