Yet Another Critique My Startup: Face Detection & Photo Organization
20 comments
Slick website. Video is very well done, one of the best "intro videos" i've seen in awhile. Was that made with Screenflow? Curious. Product seems pretty cool, I wonder can I tag other images, not just faces?
Nice job man :)
I'm working on the same project too. But my focus has been on Recognition more. Detection is a good approach to finding faces in photos (and hence faster queries), but you will understand that if you want to be successful in real world, you HAVE to extend your software to do more amazing thing :)
So:
1- Extend your system for Linux and Windows platforms as soon as possible.
2- Start to stretch its capabilities to recognize faces too.
Don't worry. You can get it done.
All The Best , Zentux (http://zentux.blogspot.com)
Don't worry. You can get it done.
All The Best , Zentux (http://zentux.blogspot.com)
Nice product. Now, stop smoking and go off the Red Bull binge so you can get your thinking cap on for the next step. Good work. I think it'll gain some traction.
Seems very cool, but the 92% of us that aren't Mac users can't really try it yet. (ok, the Mac % is probably higher on HN, but still ...)
Love your home page. Very nicely laid out, clean design. You do a great job of presenting all of the major facts in an easily-digestible format. Also, embedding the video right in the page made it much more attractive for me to click. I'm a Mac user, so I'll have to give it a whirl when I get more time. My initial impression is very positive, though.
You have to run with it, picasa 3 beta has face recognition backed into it along with googles money!
I don't think Picasa 3 desktop software has face recognition. The new version of picasa Web however does. That seems to be the common misconception between the two product updates that launched at the same time. I'm sure recognition is in the google works for the desktop app though.
I have picassa 3 and it has a face recognition tab. use it everyday.
Ahh, didn't realize that. Thanks for letting me know.
I am always thinking about the idea that one can upload a face image, then you present all the images on the web that have the face on them. Wow, that's a face search engine. Can your technology do this?
Make the client in something cross platform - you might want to try Scala, which compiles to Java byte code but suck less - an OSX client is a big turnoff, at least for me.
Never used Scala. Qt rocks though.
Originally when we were talking about building it on other platforms, Qt was the main consideration.
I love the idea, but it didn't recognize that many of the faces in my photos.
Yeah, it's not the greatest technology but we're working real hard to improve it and have made great strides already.
One of our biggest interface components is our tagger where you can get through your photos pretty quick adding and manipulating boxes on faces.
One of our biggest interface components is our tagger where you can get through your photos pretty quick adding and manipulating boxes on faces.
We finally, sort of, got our app out the door and launched in the wee hours of the morning.
>>> http://ilovephotos.com
Sadly we sort of went the "Hollywood" launch and "released" at Demofall 2008 last week... when our stuff wasn't finished. I can't count on the hands in one elementary school the number of red bulls and cigarettes (non smoker) that I have consumed in the past week or two. Too many.
Anyway, we've put it out in the wild and I would love to hear any sort of feedback you guys might have on the website and our desktop app. SORRY! but our app is OS X only at the moment. We do have plans to go onto other platforms, but for now the rest of the world is stuck using our online lightweight version of the app, which is a flash app on our website.
It's an app created to make organizing and sharing photos easier. We use facial detection (not recognition, we just find faces, or try too =]) to make tagging your photos faster and more fun. Everything is a "visual tag" rather than just a bit of text. A tag doesn't even need to have text associated with it.
We hope to leverage the data on who is in each photo do do a lot of various things. Slideshows become smarter, the most obvious improvement being with a ken burns type of effect the faces in the photo are the focus rather than getting cut off. We can also observe relationships created by photos, etc... some pretty cool stuff. Sharing becomes easier too because you can simply, in a native drag-drop manner, select some photos and drag in some people to share those with, click a button, and the rest is history.
All you mac users out there, give it a shot! I hope you like it! It's certainly beta, it certainly has some problems, but we think it's in a pretty good place right now :)
Direct download for the lazy is -> http://static.ilovephotos.com/download/iLovePhotos.dmg
P.S. It's all Django on the back end :) Nginx servin' it up.