Sorryifittakesasecondtoload,Heroku'sFreeTierputswebappstosleepafteranhourofinactivity
-------------------------------------- I've always been annoyed with landmarking images, or just getting (x,y) coordinates from an image; so I've been building up this tool to just aid me in the process. The original idea is that you can classify subsections of points by Shape.
Say, in my demo's case Nicholas Cage's Left Eye.
- You click "New Shape", label it ("Left Eye"), and the shape is now active.
- Clicking on the image adds a landmark point and stores the actual coordinates for exporting.
- You can adjust any point by clicking and dragging it
- You can delete a point by dragging it to the trash bin
- Rinse and repeat until done, Export
Your active shape expands because I want to add in some extra customization (colors, etc), I'm just drawing a blank on what else to add.
In the future, I want to add user accounts, so you can store your landmark points, and offer to build models for commonly labeled objects (faces, animals, whatever). Also, having an API for retrieving your points for any/all images in your account and finally autodetecting with any models you build.
-------------------------------------- If you were to use this tool, what would be some things you'd expect to be able to do?
If you already use something to landmark images, what? I'd love to see what they do.
Do you think understanding how to move the image after making a shape is intuitive enough, or should I add something? (Have no shapes active).
Thanks!
[link][comment]