[alpha 1] Beginning thoughts…

I uploaded a pretty large set of photos to the RIya alpha. They say it works better if you have 1,000 photos, but I don’t have that many stored on my computer right now, so I uploaded everything I could. The process took about 45 minutes, and then the server emailed me at my registered account to let me know it was done uploading. It’s supposed to show you a list of faces that the software has recognized that need tagging, but all I got was broken image links. I expect quite a bit of that; Tara made sure to let all of us testers know that the software was in “soalphaithurts” phase, but also stressed the importance of leaving feedback for the team so they can grow out of these rough spots.

I logged in to the software and started adding tags to faces. It instantly grouped together five of my pictures of Abbey, mostly because those were the shots with the best frontal picture. Riya definitely works better when the subject is facing towards the camera, but it *does* work in other situations too. It instantly made me about ten more groups of people, and I spent 10 minutes tagging everyone. The software is amazingly easy to use and runs quite smooth on my Firefox-assisted laptop here. Granted, I’ve got a cable modem all to myself, but this thing is humming along quite nicely.

After naming all my Riya-recognized people, I went through and started manually tagging people. While you’re doing this, the software is still “learning” people and their faces, so it might go ahead of you and recognize someone that you would have otherwise had to tag manually.

I’m going to download all my Flickr photos and add them to Riya tonight at some point. Right now, I’m duly impressed with what I’ve seen, even with the broken image links.

Tags: Riya, Photosearch, riyarocks