2D to 3D Image Conversions: The Process


Here is an example of how a 2D image is broken up into layers for conversion to 3D. Originally composited in Photoshop and flattened. This particular 2D image lent itself to the creation of 3 separate layers. After carefully selecting the  camera and hand, it was pasted to a new layer. The same was done with the bridge-like structure with the rails and chains. Next, the areas where the hand holding the camera and the railing have been removed, need to be filled in by cloning nearby areas on the background image, especially on the left and right sides.

In the screenshot to the right, you can see the camera/hand is now a layer , the railing and chains are also a layer and at the bottom we now have a background with the areas where all the above had been removed has been cloned to fill in. The upper layers will move relative to the background so we don’t want to see those cutouts when that happens.

Up to this point, this is typically how the 2D to 3D conversion process is done by nearly all others. A number of flat 2D layers are moved relative to one another to create the illusion of depth. You can now do this in the newer versions of Photoshop.

Screenshot of Photoshop Layers Palette

3D Wizardry takes it a few steps further. Our specialized software transforms those flat layers into layers that have their own 3D depth. Note the grayscale image layers. These are tied to the color image layers by naming with the suffix MAP. White areas are closer to you and black areas are further.  Through the careful crafting of these grayscale maps, roundness and depth can be added to each layer. A ground plane that goes from the foreground to the background can be created using a simple white to black gradient, thus one layer can have continuous depth from the foreground to the background.

Above is a screenshot from the Virtua3D module of our conversion software, Lenticular Effects. Note that the layers with a grayscale map associated have both a min and a max number ranging from 0 (background) to 100 (foreground). By adjusting these values the layer can be positioned in 3D space relative to the other layers. The layer “backgroundbackup” does not have a graymap, so only has one value. There are controls to adjust the total displacement (how far objects move), where in the depth of the  scene, the center point is to be located (this area will be sharpest because it doesn’t move), and the number of frames or views to be created.

When all this is done, applying the changes will update the scrollable animated view in the preview window to the right. When the adjustment looks good, we generate a series of  “stereo views or frames” which are then interlaced to create one very large file which is to be printed, aligned and laminated with a lenticular lens.  The result is a 3D print or transparency that can be viewed without stereo glasses. See the animation at the top of the page.

This example is a moderately complex conversion, some images can be converted with less layers, some require more. This conversion took a several of hours. Cutting out the holes between the chain links took quite awhile.  Portraits on the other hand are relatively simple. Give us a call to receive a free quote on converting and printing your 2D graphics/photos in 3D! They make great, traffic stopping signs, as well!

It may take a few moments for all the frames of this animation to load.

This is the sequence of stereo images our conversion software created from the original 2D image.

    phone: 585-278-4045              pjsucy@3dwizardry.com