The new AI experiment, CaptionBot, is self-described as able understand the content of any image and to give accurate description as well as any human.
CaptionBot can't seem to do its job pretty well at the moment and admittedly, it concedes still learning so will hold onto your photo, though no personal information is involved.
CaptionBot utilizes Computer Vision API to identify the components of the photo, and with data from the Bing Image API, it runs through Emotion API to spot the image description.
Albeit, it is able to recognise “a broad range of visual concepts” and also perform entity extraction so that it can recognize celebrities, a whole lots of work is needed to get it perfected.
Sign up here with your email