Image Detection - I want it to look in my library (bucket) only and identify based on images in my own bucket

[Dzaner Bojic]

Can Google Vision API do the following?

I want to upload my own images as the library and assign a value for what each one is.  

Then when I use the Image Detection, I want it to look in my library (bucket) only and identify based on images in my own bucket.  That is it.

Seems like it is using the whole WWW and returning all images via label detection, but I want to just use my own bucket.

If you know how, please let me know!

Please help!

[Ani Hatzis]

Yes, the Google Vision API can be integrated with your Google Cloud Storage bucket (it's also mentioned under "Integrated REST API" in the Product Overview).

There is a How-to guide that explains that you can provide an image with the request by

• uploading the image (as a base64-encoded string)
  
• passing the Cloud Storage URI of an image in a bucket
• or passing a publicly-accessible HTTP/HTTPS URL of an image somewhere on the WWW

You can use many different clients to send the request to the Vision API, e.g. a Python app that is hosted on App Engine could send the request to the API and store the response in Cloud Datastore (one entity for each image in the bucket). Since you want to detect only images that are stored in a bucket, you could also use a Cloud Function (JavaScript) that is automatically triggered every time a new image is created in a bucket (tutorial with code sample here).

--
You received this message because you are subscribed to the Google Groups "Google App Engine" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-appengi...@googlegroups.com.
To post to this group, send email to google-a...@googlegroups.com.
Visit this group at https://groups.google.com/group/google-appengine.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-appengine/40f92121-11f3-4b16-80dc-05abdd4eb7ff%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[barryhunter]

Technically no. You can't upload your own training dataset, and train the AI to produce your labels. 

Would need to deploy your own implementation - possibly via https://opensource.google.com/projects/tensorflow - deployed on GCE

https://www.tensorflow.org/tutorials/image_recognition

https://cloud.google.com/solutions/running-distributed-tensorflow-on-compute-engine

... but have had reasonable success building a second layer on top of the API, to produce custom labels. Not sure what the proper name is, some sort of Meta-AI.

Basically

1) Run all your existing images though the Cloud Vision API, and gather the feature labels it produces. 

2) You now have a mapping of 'labels' to your own 'value' - although somewhat fuzzy. Will have a long list of labels for each value. Store these in a database

3) To 'search' images by your "value". first look at the image(s) with that value, and gather all the labels. THen look for the images with the most similar labels. 

4) Can also do 'vision lookup' by uploading a image to the Cloud Vision API, and retrieveing a list of labels. Then search your database looking for the images with the most 'similar' labels. Take the 'value from the best match(es).

It doesnt matter that what the labels Google AI is detecting on the images, even if nothing like your 'values', just that they similar 'looking'. Say you have a value of 'animal'. The vision api may detect those images as a mixture of differnt animals (cow/dog etc), but you have a mapping back and forth, via your 'training' dataset.  

Not an AI expert, so probably messing up the terminology, but the labels that Google's AI produces become almost like a set of hidden neurons on your final Meta-AI

[bFlood]

only in alpha now but isn't this what AutoML is suppose to solve? (at least the photo based portion, they mentioned there will be other AutoML modules). You load your own images, label them and then use the Vision API to access the results

https://cloud.google.com/automl/

[Barry Hunter]

Ohh, not seen AutoML before. But yes, that does look like custom trained AIs hosted in the cloud :)

Interesting. 

--
You received this message because you are subscribed to the Google Groups "Google App Engine" group.

To unsubscribe from this group and stop receiving emails from it, send an email to google-appengine+unsubscribe@googlegroups.com.
To post to this group, send email to google-appengine@googlegroups.com.

Visit this group at https://groups.google.com/group/google-appengine.

To view this discussion on the web visit https://groups.google.com/d/msgid/google-appengine/9638c9f7-63d8-457b-a047-4752054fb96b%40googlegroups.com.