Notes on image annotation · HackerPapers discussions

Notes on image annotation — HN submission

Tue, 02 Aug 2022 09:19:02 GMT

Notes on Image Annotation (2012)

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Notes on image annotation — HN submission

Mon, 03 Jan 2022 16:44:23 GMT

Notes on Image Annotation (2012)

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Notes on image annotation — HN submission

Tue, 20 Apr 2021 08:01:10 GMT

Notes on Image Annotation (2012)

2 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Notes on image annotation — HN submission

Mon, 13 Jan 2020 19:32:53 GMT

Notes on Image Annotation (2012)

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Notes on image annotation — HN submission

Mon, 10 Jun 2019 16:42:50 GMT

Notes on Image Annotation (2012)

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Notes on image annotation — mentioned by joshvm

Mon, 10 Jun 2019 03:33:36 GMT

There's a great paper about image labelling, written by an expert labeller: https://arxiv.org/abs/1210.3448 One interesting side to this is citizen science; websites like Zooniverse where you can get random members of the public to do your labelling. Technically there's little difference to what Scale does, just that it might be pictures of Pelicans or galaxies rather than cars on the road. In my experience what you're paying for with Scale is consistency. Zooniverse relies on the public being... Gaussian in quality - we needed something like twenty annotations per image I think. People spend…

Paper and discussion history. A mention is not an endorsement.