Turbo Image Filter

The Turbo Image Filter experiment leverages the Google Vision API and models to filter through images looking for objects of interest, enabling an accelerated and cost-effective labeling process.

To use the Turbo Image Filter, you provide a collection of images, some of which include an object of interest, and a specification that identifies the object of interest (its Knowledge Graph machine-generated ID or a sample reference image of the object). The product documentation includes examples of both types of specification. The Turbo Image Filter returns a Boolean indicating whether an image contains the specified object or not.

Apply for access Private documentation

Intended use

This experiment is designed to help you filter a collection of images by the objects contained in the images. A prototypical use case is to accelerate the process of labeling the images by removing those images that don't include an object of interest. Other use cases also exist, such as creating or searching archives. These use cases are not constrained to specific industries or functions.

What data do I need?

The Turbo Image Filter is likely to be effective with natural image (that is, those captured by a camera in the real world. It may not be effective with highly unusual image types, such as specialized medical images and scans.


  • Images may be stored in JPEG or PNG format, with a maximum size of 2000x2000 pixels
  • Object specifications can be a Knowledge Graph ID as a string or a reference image

What skills do I need?

As with any AI Workshop experiment, successful users are likely to be savvy with core AI concepts and skills in order to both deploy the experiment and interact with our AI researchers and engineers.

In particular, users of this experiment should be familiar with accessing Google APIs.