Launch of our public Object Recognition page

We are pleased to announce our public and extremely easy to use object recognition page. Simply drag and drop a photo onto the page and it will show the objects it finds there. There is plenty of talk about using AI to do object recognition and now you can explore it yourself! Try it here:

For organizations with a large library of images, the ability to call up all images that feature a bridge, a smile, skiers from a company outing, or any other object can save a vast amount of time when searching for suitable images. The system as it is does provides advanced automated and manual features to perform tagging of images, with DBGallery always looking to add new features in future iterations. For instance, the ability to exclude certain objects from the tagging is an identified future feature. Here, excluding tags you know you will never be interested in, such as ‘Human’ or ‘Person’ would clean up the object tagging in many situations.

The system is in the process of supporting custom models, by feeding images of a particular object through the system it can be trained to recognize specific examples. Instead of recognizing just ‘clouds’ for instance, the system could be trained to identify cloud types. Same for types of cells, or architectural styles, etc. This allows recognition of industry-specific objects and unique items. While destined for future iterations, this type of exact identification could extend to facial recognition for identifying individual people.

Talking about the object recognition solution and the freely available demo page, Glenn Rogers, DBGallery Product Manager, noted, “The ability to identify and tag each image with its content in this way will revolutionize how photo databases are used. By allowing a simple drag and drop free demonstration, we not only give photographers a valuable tool, but the more images it analyses, the better it works.”


Behind the scenes, DBGallery currently integrates with Amazon AWS’s Rekognition technology to provide object recognition.