I suggest you ...

Cluster images in series in the playlist

Related to your ranking algorithm, it would be useful to do some clustering. You might even consider (in some cases) displaying this info on screen. EG "1 of 8 similar" or "1 of 3 similar". This will be important when you have the thumbs up / Thumbs down / delete button on the phone. Sometimes I see a fresh image that is kind of a crappy shot, and I'm pretty sure that there are other less crappy shots from the same time (I shot 10 in a row and this is 1 of them). If I *knew* that was correct I'd thumbs down / X it -- but I'm not sure... so I can't do anything.

Related point: As you do more and more image processing, you can do this on the front end. EG do clustering within a short time context (60 seconds, 5 minutes, whatever) look for sharp edges (a classic IR task) and then figure out how to prompt or present to users so they can say, "this is the good one, don't show the other 5".

1 vote
Vote
Sign in
(thinking…)
Sign in with: facebook google
Signed in as (Sign out)
You have left! (?) (thinking…)
AdminAndy Jagoe (Founder, Fireside) shared this idea  ·   ·  Flag idea as inappropriate…  ·  Admin →

1 comment

Sign in
(thinking…)
Sign in with: facebook google
Signed in as (Sign out)
Submitting...
  • Anonymous commented  ·   ·  Flag as inappropriate

    Another thought on the product side. Might be interesting to experiment with photo views where sequential shots that were take at the same location are shown as a short "stop-motion" movie. Wouldn't actually be "animated" as you can't assume the pictures are very similar, but the idea would be to explore a memory a bit deeper than one random picture. So maybe take 10 photos at random from a given location and show them for 5 seconds each. One minute per "memory".

Feedback and Knowledge Base