Forgetting to unlock the face, Indian researchers have come up with a way to authenticate mobile phones with teeth

A team of researchers in India has developed what appears to be the first human authentication via teeth for mobile and other handheld devices. According to the team, the app uses a mobile phone camera to get a biometric sample. This app has specific markers that register human teeth for authentication, similar to existing apps that record the entire face.

The name of the study is “Deepteeth: Teeth for Mobile and Handheld Devices-Photo-Based Human Authentication System”. It was created by Geetika Arora, Rohit K Bharadwaj, and Kamlesh Tiwari of the Birla Institute of Technology (BITS) in Pilani.Explain the functions of apps and teams Summary of the treatiseShows that the region of interest (RoI) is extracted using markers and the obtained sample is enhanced using contrast limiting adaptive histogram equalization (CLAHE) to improve visual clarity. ..

The team states that, to the best of their understanding, this was the first work on photo-based authentication of teeth on any mobile device, and adding results showed “perfect accuracy.” If you read this treatise further, you will find a diagram that illustrates how tooth photo authentication works. This app uses the front camera of your mobile device to get the impression of your teeth first. This is followed by ROI extraction and enhancement. The next function of the app is “detailed feature extraction” followed by “registration / verification and identification”.

The next step is where authentication actually begins. The registered tooth extraction compares the impression of the tooth with the database, and then the app “determines” whether it matches the right person.

In conclusion, the authors write that less-explored tooth photographs have been observed to have very high cognitive and discriminating accuracy due to the special features proposed in the study.

Also, the initial training takes some time, but when deployed, it is very efficient for identification or verification. Studies have shown that the proposed model works perfectly with small sized samples, making it power efficient and suitable for mobile devices.

“We also proposed a new way to regularize deep learning architectures by combining margins and mutual information in backbone feature representations,” the researchers write in their research. Forgetting to unlock the face, Indian researchers have come up with a way to authenticate mobile phones with teeth

Back to top button