Apple’s wasting no time in taking strides towards technological improvement, and in lieu of that it now seeks to improve its facial recognition technology. Now, instead of just recognising facial features, Apple products will be taking body postures and other minute details into account.
Apple’s facial recognition software, labelled Face ID, launched with the iPhone X in 2017 to generally positive reception. The usual skepticism aside, it really seemed like futuristic tech come straight out of an 80’s sci-fi movie a la Minority Report (as opposed to the rather disappointing fad of hover boards). The technology behind it is no small matter either. Utilising infrared sensors that can not only visualise surface area but depth as well, it could very accurately scan people’s faces, down to minuscule pores in one’s skin surface. Such technology had been implemented in motion sensors such as those of the Xbox Kinect or the PS Move, but never to this degree. The sensor would go as far as to simply not unlock phones until the person was directly staring at the camera, to counter accidental unlocking. A rather straightforward concept, utilised beautifully. The specificity of the sensors and the security they offer is another marvel. This author would very highly recommend a video by the YouTube channel Corridor, labelled “Trying To Hack iPhone Face ID”, which goes into much more depth about the technology, how it works, and to what parameters it can be stretched.
Recognising mannerisms and poses is a slightly different ball game, and would be more effectively utilised in iPhoto, which has its own facial recognition software (which of course, is much less accurate than Face ID). While Apple itself has been tight-lipped about any such developments, it has also recently filed a patent application for the concept. The patent also delves into some detail about how the concept will work. Apparently, by going through different images of subjects, in singular or group photos, the app will then separate them into “clusters”, denoting facial features, body mannerisms, etc. As more and more photos are fed into it, the app’s AI will get more and more accurate at recognition since it now has more information to work with. Theoretically, this could get to a state where the app will be able to recognise people with masks on as well, depending on their eyebrow movements or the pose they take during a selfie. Which, even as the COVID-19 pandemic inches closer to winding up with vaccines in development, will be a relief. So fire up your phone and get to taking photos with you and your friends! The app will thank you for it.
anouchka via Getty Images
Read next: Apple Store’s Best of 2020 - How has 2020 Impacted our Screen-time?
Apple’s facial recognition software, labelled Face ID, launched with the iPhone X in 2017 to generally positive reception. The usual skepticism aside, it really seemed like futuristic tech come straight out of an 80’s sci-fi movie a la Minority Report (as opposed to the rather disappointing fad of hover boards). The technology behind it is no small matter either. Utilising infrared sensors that can not only visualise surface area but depth as well, it could very accurately scan people’s faces, down to minuscule pores in one’s skin surface. Such technology had been implemented in motion sensors such as those of the Xbox Kinect or the PS Move, but never to this degree. The sensor would go as far as to simply not unlock phones until the person was directly staring at the camera, to counter accidental unlocking. A rather straightforward concept, utilised beautifully. The specificity of the sensors and the security they offer is another marvel. This author would very highly recommend a video by the YouTube channel Corridor, labelled “Trying To Hack iPhone Face ID”, which goes into much more depth about the technology, how it works, and to what parameters it can be stretched.
Recognising mannerisms and poses is a slightly different ball game, and would be more effectively utilised in iPhoto, which has its own facial recognition software (which of course, is much less accurate than Face ID). While Apple itself has been tight-lipped about any such developments, it has also recently filed a patent application for the concept. The patent also delves into some detail about how the concept will work. Apparently, by going through different images of subjects, in singular or group photos, the app will then separate them into “clusters”, denoting facial features, body mannerisms, etc. As more and more photos are fed into it, the app’s AI will get more and more accurate at recognition since it now has more information to work with. Theoretically, this could get to a state where the app will be able to recognise people with masks on as well, depending on their eyebrow movements or the pose they take during a selfie. Which, even as the COVID-19 pandemic inches closer to winding up with vaccines in development, will be a relief. So fire up your phone and get to taking photos with you and your friends! The app will thank you for it.
anouchka via Getty Images
Read next: Apple Store’s Best of 2020 - How has 2020 Impacted our Screen-time?
