Swapping faces
Digital surgery creates the perfect expression from multiple images
By Stephen Romaniello
Story and images by Stephen Romaniello
Lustre Art Group – stock.adobe.com
Sometimes the photo just isn’t right. Perhaps it’s the wrong expression or the subject blinked at exactly the wrong moment, or the smile looks more like a grimace. Even worse, an important person showed up too late to be part of the group photo. These are common occurrences that could spell catastrophe for a photographer or digital artist with a critical deadline. The good news is that with a few precautions and a little digital magic, these potential disasters can be corrected. This article will take you through the process of face swapping and body inserting. No, this is not a creepy science fiction B movie; this is about using sophisticated editing techniques for seamless compositing of image content from multiple sources.
Shotgun
The process of compositing starts at the photo shoot. Professional photographers shoot dozens of photos of the same subject. Especially for portraits and group shots. Shooting multiple shots is sometimes referred to as the “shotgun technique” and practically guarantees success. If the “money shot” doesn’t happen in the first dozen, it’s more likely to happen in the second or third dozen (Fig. 1). Even if there is no perfect shot, elements from multiple images can be combined to produce a perfect composite. Think of it like shooting a shotgun. The more pellets that fly out of the barrel, the more likely you are to hit your target.
When you take the photograph, work the scene. Engage with the subjects and try to make eye contact. Why? Because, if you engage with your subject, you can get them to relax. And if they are relaxed, the more likely you are to capture a great image. Sometimes a subtle difference between what is happening in the background, the eye contact of a person, or a hand gesture is what makes the photograph.
Swapping features
Sometimes it comes down to a simple replacement of a sad or open mouth or closed eyes with open ones as in the case of (Fig. 2). This photo was taken during a portrait session, and fortunately, the photographer snapped dozens more, because it was later needed for publication in a newsletter with a tight deadline. Ultimately, three photos were combined to produce the perfect portrait.
The source images were chosen because the features were good and the poses were similar to the destination image. The eyes and mouth are selected on each photo using the lasso tool, leaving some space around the areas including the eyebrows and the mustache — which can always be masked out if they are not needed. The selection marque is feathered (Select > Modify > Feather) in order to create a soft transition zone (Fig. 3).
Figure 1. If the “money shot” doesn’t happen in the first dozen, it’s more likely to happen in the second or third.
Figure 2. Sometimes it comes down to a simple replacement of a sad or open mouth or closed eyes with open ones.
Eye-mouth coordination
Drag and drop the selected areas, on each source photo, one by one, to the destination document and roughly position them over the underlying eyes and mouth. They appear on separate layers which will be named. Reduce the opacity of the layers to 50% so that the underlying content can be seen to better position and scale them. They can be sized or slightly distorted, (Edit > Transform > Distort) so that can be perfectly aligned with the underlying content. After they are properly in place, increase the opacity to 100%.
Figure 3. Use the lasso tool to encircle areas on the target image. Feather the selection so that it has a soft transition edge.
Figure 4. It’s often necessary to perform a color adjustment to better match the color and tonality of the original.
Color adjustment
It’s often necessary to perform a color adjustment to better match the color and tonality of the original (Fig.4). Use a Hue/Saturation, Levels, or Curves adjustment layer to match the new layer content to the destination image. The two adjustment layers are “clipped” to the Eye layer and Mouth layers independently so that they only affect the color of the layer immediately below them for more precise control. To Clip Layers, place the cursor between the layers in the Layers panel. Press Option (Mac) or Alt (Win) and click.
Missing persons
A group shot can present difficulties if an important individual, for one reason or other, cannot attend the initial photo shoot. They’ll need to be inserted later, and that requires a separate photo shoot. If the initial group shot was taken outdoors, the rescheduled shot should be taken in the same location, at the same time of day, and under similar lighting conditions. If that is not practical, which it often isn’t, care should be taken in using Photoshop’s adjustment layers and filters to best match lighting conditions and tonal qualities of the original image.
Compositing an outdoor group has its own unique set of challenges. If you can decide where the missing person will be inserted beforehand, at the initial shoot, then you can plan to leave a space for them. More frequently, however, you don’t have that luxury.
It’s easier to place the missing individual in front of the group, but sometimes that can appear awkward. Closer to the back requires that you edit content by masking. Finding a simple background for your subject will make compositing easier.
The original photo without the missing person is shown in (Fig. 5). You can see where the photographer left a space on the right to place him. The results of the second photo shoot (Fig. 6), show where the subject is standing in the location where he will be placed. The work of compositing has been substantially simplified by shooting against a relatively simple background — the green door, the terrazzo tiles, and the white column. The new content layer with the missing person was dragged and dropped into the document and the new layer’s opacity was reduced to 50% so that the figure can be sized and accurately placed (Fig. 7). A carefully-made layer mask was traced to eliminate unwanted content. Finally, the new layer’s opacity is increased to 100% to complete the composite (Fig. 8).
Figure 5. The original group photo without the missing person.
Figure 6. The results of the second photo shoot, where the subject is standing in the location where he will be placed.
Figure 7. The new content layer with the missing person was dragged and dropped into the document and the new layer’s opacity was reduced to 50% so that the figure can be sized and accurately placed.
Figure 8. An accurate layer mask was traced to eliminate unwanted content. Finally, the new layer’s opacity is increased to 100% to complete the composite.
Shadows
One of the most the most common mistakes when compositing an image is that the new content seems separate, does not look anchored to the ground, and appears as if it’s been stuck to the original. An easy way to avoid this problem is to select enough of the background where the figure stands so that the natural shadow is included.
If you’re unable to select the background and need to make a tighter selection of the figure, you can add a slight shadow with the burn tool to darken the areas that contact the ground (Fig. 9). It’s a quick fix, but it works to better anchor the new content to the image.
Figure 9. Add a slight shadow with the burn tool to darken the areas that contact the ground.
Compositing tips
- Take care when choosing content. Be sure that the light sources are similar in both images.
- Pay attention to the details when selecting content. Include everything that is relevant in your initial selection and then some. It’s much easier to eliminate content by masking than it is to include elements that you neglected.
- Feather the edges of your selections for better transitions.
- Use a soft-edge brush on your layer masks and reduce the size and opacity of the brush to blend adjacent areas.
- Correct the color and tonality of the new content so that it perfectly matches the original image.
The techniques demonstrated here are standard practice for salvaging images. I’m sure you’ll find them useful. Compositing can be a lifesaver and sooner or later every professional needs to accurately combine the content of two or more images. Of course, it’s always better to get the “money shot” the first time, so shoot multiple images of the same subject. Even then, it may be necessary to composite elements from multiple sources.
Stephen Romaniello is an artist and educator who has taught graphics software for over 30 years. He is a certified instructor in several graphics and video applications and the author of numerous books and articles on the creative image manipulation, illustration and media programs. Stephen is the founder and CEO of GlobalEye Systems, a company that offers training and consulting in graphics software. He is also the founder of Fire Boy Productions, an art-book publishing company.
