Description text

Each picture is analyzed for content that corresponds to a tag. The colors in the bottom picture show the likelihood that the object in the picture corresponds to a specific tag.