Hi everyone, I was creating a background image for an assignment at work, and a friend of mine showed me a very cool tool in photoshop cs3 for stitching panoramic images. Anyway, although its designed for panoramas, being that Im into creating characters I decided to try this tool to create a face texture and ended up with some pretty nice results in with about 10 minutes worth of work. So I decided to create a quick tutorial on it you guys and gals might find helpful. Big shout out to cooperunionstud for showing me this tool which allowed me to even think about trying something like this.
You will need a set of images from the different angles that you want to blend together
So Step 1: Go to File/Automate/Photomerge
Step2: Make sure you select the interactive layout option so that you can position your images properly. click browse and select the images you wold like to use..
Step3: The top bar will have thumbnails of your selected images which you can drag onto your work area. You might want to unclick snap to image over on the right side. Try to position the images by lining up key details such as the corners of the eyes and mouth. You may need to open the images separately and use the transform deformation tool to tweak the images to line up a bit more prior to your importing them into photomerge. Once your satisfied with the alignment, click ok...
Step 4: if all goes well Voila, you should have something similar to this and you can now use any bruch to adjust your blending on the mask. each piece will have its own layer and can be adjusted separately.
Now you can adjust your uvw to match this or us Photoshops Transform deformation lattice or liquify to fit this to your uvw pretty quick. I hope this is helpful to you guys...
Later
Dennis
Replies
I wish I would have know this last semester when I was modeling heads for school would have saved a lot of time
Hi Dennis,
I am by no means a character modeler but I know that our folks here at EA have a very nice method to very quickly texture their faces.
Once they have unwrapped their model on channel 1, they are going to add another UVW map modifier to their stack on channel 2 and project a photo of a face (the side for example). They will then add a UVW unwrap modifier on channel 2 and tweak the vertices to match the model itself (it's basically stretching the picture, but since it's so high-res it doesn't really matter at this point since it will be rescaled.)
They will then render to texture the result on channel 1, take this to photoshop, repeat for the front and the back of the face then erase the seams.
It's really similar to your way of doing it, except it's the other way around. Just a different method of working.