Creating a Composite Picture to Make a Post-Apocalyptic Survivor

Post-Apocalyptic Survivor

This project was to practice combining two different pictures together to make them look like one. During the Covid "apocalypse", I took a picture of myself wearing a fancy portable air-filtration system. I also added my swim goggles for extra effect.  Not wanting any good picture go to waste, and to try something different than natural landscapes, I thought it would be fun to take a go at a science-fiction theme. 

Original

In Affinity Photo 2 (AP2) I used the Selection Brush tool with "snap to edges" and "soft edges" options and selected the background. The soft edges option reduces the hard "cut" that looks jagged and unnatural when placed against a different background.  Under the Layer option, I selected "Inverse pixel selection", by doing that it selected me, the post-apocalyptic survivor.  I created a second layer and copied and pasted the selection. Finally, I returned to the background and erased all remnants of the kitchen wall.

Next, I did a Google picture search for an apocalyptic background I could use.  I saved one and used "File--Place" within AP2 to place the picture in the scene.

Original Background
https://www.peakpx.com/en/hd-wallpaper-desktop-fiywo 

The background scene was much larger than my photo so I moved it to get a good background image.  Then I saved the picture as a JPG and opened it up in Luminar NEO for the rest of the project.  There was still much work in order to make it look like I belonged in the picture.  That's where the magic of NEO came into play.

My first task in NEO was to darken the survivor. The original lighting made it look like there was a flash. In the scene, the sun is behind me, and naturally, my frontside would be darkened.  So in the Edit menu, I used the Develop option to reduce the exposure by 2.35.  To do that I selected Mask, then Mask AI. At that point, the NEO AI analyzes the photo and identifies the different features of the photo.  That could be sky, landscape, buildings, people, faces, etc...The first time around it did not identify my body and head. Not too surprising since there is a contraption attached to it and I am wearing goggles. So I canceled the operation and tried again. BINGO!  I was happily surprised a different AI analysis got better results.  

After darkening the exposure, I adjusted the hue to a more yellow to try to match the background.

Next, I added a texture (acrylic finish) layer on top of the picture. This accomplished two things. First, it gave the picture more of an old-picture feel. But most importantly, it really merged the two pictures together, in both lighting and texture.  

Next, I used the Atmosphere edit to add a little fog in the midground.  That again increased blending and diminished the acrylic finish texture across my forehead. 

To give it a more gritty feel, I used the Mood option and add "Color Punch Hot" at 97%. I played with this a lot and almost didn't use it.  

I used "Dodge and Burn" to lighten and draw out the goggle lens white areas. If you go to the original, you'll see what I mean.  Before, the lower contrast made the goggle lenses all black, not enough grittiness for me. I also used "Dodge and Burn" to darken the shadows of the mask and hose.

In my finishing touches, I used the cloning tool around the perimeter of my body to reduce a slight yellow hue, mostly around the shoulder.

Overall, I was pleased with the final product.  It was a challenge I wasn't sure I would be able to pull off.  I really sharpened my editing skills and confidence on this one.  I look like a legit Post Apocalyptic Survivor!