5 ms·

How would this work for image and facial recognition, especially with regard to individuals? It seems the features needed to recognize a face are by definition

How would this work for image and facial recognition, especially with regard to individuals? It seems the features needed to recognize a face are by definition identifying.

samch 10y ago
If you're referring to the new facial recognition system in iOS 10's Photos app, I believe they mentioned that the entire system would run locally on your own hardware. I took that to mean that they would not be sending anything, identifying or not, to Apple's servers.
NovaS1X 10y ago
From what I understand from the conference is that facial recognition is done locally on device and not transmitted. They said that quite clearly. There is probably some specifics I'm missing though, and it raises the question about how they identify others faces. I've not used the feature so I can't say. I'm guessing it's a manual tagging when the device can't find a match and then on device processing afterwards once the algorithm has learned.
curt15 10y ago
Google photos only groups similar faces and requires the user to assign a name to that group. Apple could presumably perform the grouping in the cloud and relegate the association of the group with an identity to the user's device.
- droopyEyelids 10y ago
  In the keynote they specifically called out face & image recognition as something performed entirely on-device, without touching the network.
  - marcuskaz 10y ago
    That was my understanding too, but seems like to handle the vast amount of image recognition required for faces and also objects which they demoed; there would need to be a large data set to learn from, and results from that learning saved. The data set to learn from probably can be gathered from stock photos or other open source images, but can that learning be saved down to phone and used to compare against, or would it be too big? If too big, does that mean characteristics of an image would be computed on the phone sent to the cloud compared and the results sent back?
    - GrantS 10y ago
      Yes, the results of all the learning can be saved in a compact form which is efficiently run on the device. Apple just yesterday unveiled their neural network API [1] which they are presumably using for these tasks. The description states: "BNNS supports implementation and operation of neural networks for inference, using input data previously derived from training. BNNS does not do training, however. Its purpose is to provide very high performance inference on already trained neural networks." So they're really saying that no data leaves the device for either object or face recognition at classification time. Now, whether they are using iCloud-stored photos (which the user has already agreed to share) for training, I don't know, and any privacy issues would still be important at that point. [1] https://developer.apple.com/reference/accelerate/1912851-bnns https://developer.apple.com/reference/accelerate/1912851-bnn...

How would this work for image and facial recognition, especially with regard to individuals? It seems the features needed to recognize a face are by definition identifying.

samch 10y ago

If you're referring to the new facial recognition system in iOS 10's Photos app, I believe they mentioned that the entire system would run locally on your own hardware. I took that to mean that they would not be sending anything, identifying or not, to Apple's servers.

NovaS1X 10y ago

From what I understand from the conference is that facial recognition is done locally on device and not transmitted. They said that quite clearly. There is probably some specifics I'm missing though, and it raises the question about how they identify others faces. I've not used the feature so I can't say. I'm guessing it's a manual tagging when the device can't find a match and then on device processing afterwards once the algorithm has learned.

curt15 10y ago

Google photos only groups similar faces and requires the user to assign a name to that group. Apple could presumably perform the grouping in the cloud and relegate the association of the group with an identity to the user's device.

droopyEyelids 10y ago
In the keynote they specifically called out face & image recognition as something performed entirely on-device, without touching the network.
- marcuskaz 10y ago
  That was my understanding too, but seems like to handle the vast amount of image recognition required for faces and also objects which they demoed; there would need to be a large data set to learn from, and results from that learning saved. The data set to learn from probably can be gathered from stock photos or other open source images, but can that learning be saved down to phone and used to compare against, or would it be too big? If too big, does that mean characteristics of an image would be computed on the phone sent to the cloud compared and the results sent back?
  - GrantS 10y ago
    Yes, the results of all the learning can be saved in a compact form which is efficiently run on the device. Apple just yesterday unveiled their neural network API [1] which they are presumably using for these tasks. The description states: "BNNS supports implementation and operation of neural networks for inference, using input data previously derived from training. BNNS does not do training, however. Its purpose is to provide very high performance inference on already trained neural networks." So they're really saying that no data leaves the device for either object or face recognition at classification time. Now, whether they are using iCloud-stored photos (which the user has already agreed to share) for training, I don't know, and any privacy issues would still be important at that point. [1] https://developer.apple.com/reference/accelerate/1912851-bnns https://developer.apple.com/reference/accelerate/1912851-bnn...