When you upload an image to a thrill post, you'll see an "Analyzing Thrill" message. During this process, referred to as "image detection," we'll scan the image content to identify things like well-known locations and common activities, as well as content that is likely to be blocked, such as excessive violence or nudity.
The ThrillPiXx application utilizes Google's Cloud Vision API (application programming interface) to scan uploaded images. Data from Google's API is then mapped to our thrill categories. If something doesn't match one of our categories, we add it to the Other category.
Image detection can be used to automatically generate hashtags and descriptions for the thrill post as well as automatically select the appropriate category. For example, if someone riding a horse is detected and the Pyramids of Giza are also detected, hashtags might include #egypt, #pyramids, or #horsebackriding, and a description might be Horseback riding in Egypt.