This is a programming sub project really. I am working on creating a web site with a SQL database behind it to document my computer game collection. One of the key differences between my projects and the many similar projects out there is that I plan to include 3D models of the game boxes.
I am creating the 3D models in Blender 3D 2.8. To get the textures, I started by taking pictures of each side of the game box and editing them into textures in Photoshop. I wanted to make it so people that didn’t have Photoshop could document their game collection and upload the models to the database.
The tool is a work in progress but here is the basic idea. You load all of the photos you took of your box and they appear on the left.
![](http://52.27.30.253/wp-content/uploads/2019/06/3DKLoadedPhotos-1024x582.jpg)
Currently the computer vision attempts to guess where the edges of the box are. It doesn’t work that great yet. That is why I have included the ability to manually edit the corners. I have included a screen shot of a case where the computer vision worked and one where it failed.
![](http://52.27.30.253/wp-content/uploads/2019/06/3DKCVDebugWindow-1024x585.jpg)
![](http://52.27.30.253/wp-content/uploads/2019/06/3DKCVDebugFail-1024x584.jpg)
Here is what editing the failed guess looks like. As you move the corner points the preview pane on the upper right updates in real time.
![](http://52.27.30.253/wp-content/uploads/2019/06/3DKManualEdit-1024x582.jpg)
![](http://52.27.30.253/wp-content/uploads/2019/06/3DKManualEditDone-1024x582.jpg)
I have more features planned and I also have a lot of work to do on the web site front end as well.