3 ms·
Project Oxford – An evolving set of APIs for natural data processing
- danso 11y agoPretty slick...the demo page is pretty fast and easy to use: https://www.projectoxford.ai/demo/visions#Analysis https://www.projectoxford.ai/demo/visions#Analysis I tried a few photos...the categorization of the photos wasn't as impressive as the examples they showcase...for example, in every photo, "outdoors" was the only tag I got, even for this photo which is inside a subway car (I was hoping it would receive the train/station tag, which is apparently one of the things it recognizes in the demo photos): http://i.imgur.com/hpBqGYw.jpg http://i.imgur.com/hpBqGYw.jpg On the other hand, it picked out 6 of the 7 faces in the photo, correctly determined their gender, and (AFAIK) gave reasonable estimates to all of their ages. http://i.imgur.com/lEugMR9.jpg http://i.imgur.com/lEugMR9.jpg
- Syrup-tan 11y agoThe ``Adult Score'' and ``Racy Score''s seem to work rather well on drawn content[0][1] [0] https://imgrush.com/26f8ec00a4b0 https://imgrush.com/26f8ec00a4b0 [1] (NSFW) https://imgrush.com/be7c4078c7a0 https://imgrush.com/be7c4078c7a0
- jonas21 11y agoInteresting. Clarifai has a similar image tagging demo (http://www.clarifai.com#demo http://www.clarifai.com#demo), and their results seem more reasonable: http://i.imgur.com/o2FVfSo.jpg http://i.imgur.com/o2FVfSo.jpg
- doppenhe 11y agointelligent services is where the future is. Being able to build these types of capabilities into any app anywhere will just empower developers to use the state of the art without having to implement from scratch. Pretty cool to see Microsoft getting onto this path that other companies have started down.