Facebook DeepFace facial recognition fail vs success

Facebook released its new research paper going into great detail about the new ‘DeepFace’ facial recognition technology, since this occurred it had had many mixed comments about it – Will the new program be a fail or a success?

The new facial recognition program created by intelligence analysts includes AI Yaniv Taigman, Ming Yang, Marc’Aurelio Ranzato and Lior Wolf; this new tech is so accurate it is almost the same as human perception.

It is accurate about 97.25% of the time, the new dataset uses 4 million images of 4,000 faces, once this has been done DeepFace will look at the angle and correct where need be on each face, the face is analyzed before comparing it to a 3D model of the average face. Once this has been done the program then assigns a numerical description that allows the face to be compared to other images.

No one knows for sure how Facebook will implement DeepFace for use as of yet, but could be used for so many different things with Facebook, a camera fitted in shops that recognises your face and will then show you on your mobile or home device product ads that you maybe interested in.

There is no release date for DeepFace yet but we will be sitting here waiting for the announcement, the research paper titled ‘Facebook Deepface facial recognition fail vs success’ goes into great detail about the new DeepFace technology and is well worth reading.

We talked to a few of our friends about Facebook’s DeepFace and they said it would be fantastic if it were open source – Do you think DeepFace should be open sourced?

DeepFace, do you think the program will fail or be a success? Please do use the comments below to have your say.

Facebook DeepFace facial recognition fail vs success