Posts Tagged ‘Photo Editing’

I kick myself every time I mess up the settings on my camera. On Nicole’s first day of school I mistakenly left my camera in full manual mode, with the shutter speed set for indoor photography of stationary objects. The photo on the left is my favorite pose wise. Nicole is so happy. But the photo is so over exposed with a slight motion blur given the slow shutter speed. Fortunately I realized my mistake, adjusted my settings and was able to also capture the photo on the right.

Current state of the art photo editing techniques cannot save the photo on the left. The skin on the left side of her face is so over exposed the sRGB value is pure white. The computer simply has no way to figure out what color was supposed to be there, so there’s no way to automatically fix the white patch. Nicole could have cyan colored skin for all the computer knows. Or checker pattern skin, for that matter.

As an machine learning person, I find this limitation incredibly irritating. Chances are, when you, a human, first looked at the image on the left Nicole’s skin did not appear pure white to you. Your brain filled in a likely color based on the right side of the photo where her skin isn’t as badly over exposed and what you know of human skin tones. With the two photos side by side you can easily imagine a combined photo with the pose from the left and the coloring on the right.

Current state of the art image recognition can recognize faces. It should be able to map point for point the location of the eyes, mouth, nose, hands, etc between the two photos. Given that, it seems plausible for a AI enabled photo editing software to merge the two photos and create the ideal image.

Even without the image on the right, a sufficiently advanced AI based system should be able to recognize the subject matter much the way a human would. The system should then be able to generate a hyper realistic plausible image. It might not be the “correct” image, the image that would have been created if my settings were correct in the first place. I doubt for most people that would matter. A realistic enough resulting image that was plausibly correct would likely be sufficient for most momtographers like me.

November 15, 2014

Half Naked Maternity Photos

Showing off the 39 week baby bump. I used the ios app Waterlogue to give it the effect of a water color painting, and to hopefully make the photo less objectionable to anyone out there who may object to half naked photos.

A little over two months ago I managed to take a pants-less maternity photo. It wasn’t intentional. I wanted to avoid the ugly full panel waste line in my photo so I opted to only wear an oversized sweater. I was using a camera tripod at the time, so I couldn’t see how I was framed before the photo, and ended up with a little bare leg. I loved it. There was something about showing a little skin that made me feel beautiful again, the kind of feeling I hadn’t really had since my last pregnancy.

39% of the maternity photos I’ve taken this time around have been sans pants.

I kept wondering what is it about being pantsless that made me feel better about my figure? If I’m pantsless you do get more of a sense of my figure. I’ve recommended form fitting clothes for maternity photography before. A while back I had read that women who gain a little weight tend to want to hide it with baggy clothes. But baggy hides all curves, which can actually give the illusion that you’re bigger than you actually are. I found that if I wore a t-shirt, it would hang off my chest or bump. You couldn’t always tell it was a baby bump under there. Coupled with baggy pants and you had no idea how big my hips and thighs were. Sometimes I just looked large. If I went sans pants in a photo, my legs looked really skinny beneath the giant belly. Or so I think, anyway.

I also found it somewhat empowering to show a little skin. I may be pregnant and have gained a bit of weight this time, but I’m still owning my figure. Confidence tends to photograph well.

The more I think about it the more it makes me a little sad that there’s such a stigma against taking even partially nude photos, even where you cannot see anything. There always seems to be the assumption that the intention is to take them for other people, to show off, or get the ‘wrong’ kind of attention. The photos I took made me feel great about myself at a time when I was otherwise self conscious. I have no intention of sharing any of them, besides the example above that has been significantly altered from it’s original form. I wouldn’t want anyone pressured out of taking a photo like this any more than I would want someone pressured into it. If your comfortable with it, and it will make you happy, go for it!

September 20, 2014

Pixel Count Convert

I find myself increasingly a pixel count convert. For years I’ve been on the megapixels-don’t-count bandwagon. My old 10.2 megapixel D60 was capable of producing an image large enough to get a 20 by 30 wall print, if but just barely. More megapixels don’t necessarily translate into a better quality image, but they do add up in terms of file size. I’m already running out of space on my hard drive. Again.

But every once in a while I start wishing I had more mega pixels.

This weekend I wanted to take more maternity photos, this time with my face showing. Nicki, who is quite the little ham these days for the camera, wanted to be a part of it. She kept running in and out of frame. In most of the photos she’s weirdly cropped, or there’s a bit of motion blur. I swear toddlers must be related to the flash. Even at 1/640 seconds, with my 50mm prime she can be a blur on camera. I still love the photos with my little toddler photo bomber (what parent wouldn’t?) but they’re not something I intended to share.

Then I came across this one:

Straight out of the camera
Straight out of the camera

After some cropping.

Cropped to capture a sweet moment
Cropped to capture a sweet moment

I love it! Since it’s a 60% crop I can get a 5 by 7 high res print easily. I can even (probably) do a 8 by 12 with a little tweaking. It’s moments like these that make me grateful my current camera is 16.2 Megapixels, and even makes me wish it was a little higher!

Photo retouching is a fact of life in the professional photography world. Just as an article wouldn’t go to print without being reviewed by an editor, a photograph isn’t published without first passing through Photoshop or equivalent editing tool. Yet, there is single level of photo retouching or manipulation that is universally considered acceptable.

Last year photo forensics experts accused the winner of the world press 2013 photo of doctoring the photo. Experts used error level analysis which led some to believe the image was a composite of multiple images. Even after the photographer later released the RAW file (a file generated directly from the camera’s image sensor), the debate continued on just how much editing was preformed. While the award wasn’t stripped the World Press decided make changes for it’s upcoming competition.  The debate continues, how much is too much manipulation?  What is the landscape of acceptable photo retouchings?

About six months ago I started conducting some unscientific polls to understand public perceptions of photo manipulations better.


Authentic Vs Idealized
The primary theme that came out of my discussions was that of authenticity or idealization. While photographers generally try to capture both truth and emotion, an authenticity view prioritizes truth over emotion, where idealized prioritizes emotion over truth.

In terms of personal photography, an idealized photo might try and showcase the best of a subject. As any teenager will tell you, some days are good acne days, others are not. An idealized view point may attempt to reduce the acne so that the photo appears as though it was taken on a good day. For those on the idealization side of the axis, this type of action is not altogether dissimilar to finding a flattering angle, removing objects out of the background, or even using makeup. Incidentally, some people on the far end of authentic would agree, and disapprove of these actions as well.

Acceptance vs Privacy
There was a secondary theme that began to emerge throughout the discussions, especially in context with online photo sharing (blogging, facebooking, instagraming, etc): that of acceptance and privacy.

The idea behind acceptance is that by display a photo ‘as is’ you are showing your acceptance of the subject matter. Someone who leans strongly into the acceptance side would argue that by editing the photo, say removing acne, you send the message that acne is bad and needs to be hidden. Such an action could effect the self esteem of not just the subject, but presumably others who may relate to the subject.

On the other side, some argue that editing a photo may allow the subject or the photographer to hide a piece of themselves they want to keep private when posted publicly, or even pseudo-publicly. For example, I can’t fathom caring if a naked baby photo of me ended up on the internet, but I would care if random strangers knew where all the freckles on my body were. For me, the baby photo isn’t an invasion of privacy, per say, but the revealing of such distinguishing marks is.

No One Right Answer:
Almost all individuals surveyed had limits on how much photo retouching was acceptable, however, those limits weren’t uniformly applied. Opinions seem to differ depending on the use case. Like all things, it depends on the context. Acceptable treatment for a photo kept privately differed from those intended for small audiences, and those showcased publicly.

Perhaps related, one key aspect in whether a treatment is considered acceptable appears to be intention behind the retouching. In the world press photo contest, and photo journalism in general, the idea is to inform. In these cases the retouchings were often viewed as deceitful because they can alter the viewer’s perception of world events. Fewer people took issue with whitening teeth, or removing wrinkles from personal photos, than from magazine editorials, even though such edits were perceived as more common in magazine editorials. No one like feeling tricked.

In our newborn care class we were instructed not to shampoo our baby in the first few weeks, to only use water. Shampoos can be abrasive on newborn delicate skin, besides newborns aren’t supposed to get very dirty. Nicki, however, was born with a full head of hair. Hair gets oily after a few days of not washing it, whether on a newborn or a full grown adult. After a few days it was quite obvious that we were sticking to the no-shampoo rule. In retrospect it was probably silly for us to have stuck so firmly to the rule. In the hospital they cleaned her with diluted soapy water right after she was born. Certainty we could have done the same on an as needed bases. But we didn’t.

I wanted to use a nine day old photo for her birth announcement, but the awkward teenager with oily hair inside me cringed. So I decided to see if there was anything that can be done in post processing to reduce the shine. Short answer: not really. You need mad fat skills to fix that level of hair problem. Still, I found a solution that kinda sorta worked. Shortly there after we started using shampoo.

Then last night I captured this.

Nicki hugging her daddy, looking at the Christmas tree.

I love it. It’s one of those moments that’s hard to manufacture. The photo was taken post bath. We pat Nicki’s hair dry, and let it air dry the rest of the way. I hadn’t gotten around to combing it yet, hence it’s still a little on the straggly side. The photo reminded me of her birth announcement photo, and decided to take another crack at smoothing her hair in post processing.

The goal is to reduce the contrast in the hair, without losing the texture. Thus the smoothing brush won’t work (it destroys the texture). The dodge (lightening) and burn (darkening) tools are risky and can make things worse. So instead I used the cloning brush.

Start by creating a second layer. I cloned her hair from the original image into the second layer. The intent is for the hair to be in a slightly different spot, so that it’s flowing in the same direction and the highlights in the hair from one layer are over/under the shadows in another. The new image should look very similar to the original.

The side by side. Left original, right with the clone. The difference I hardly noticeable, even side by side.

Next I reduce the opacity of the second layer to about 40-50%. If the cloned region is in good spot, the texture remains the same, but the overall lightness/darkness is more even. This gives the appearance of slightly smoother hair. As a final touch I use the eraser and touch up the edges of the cloned layer so they’re not noticeable.

Don’t worry about light region over top of a light region, or dark over top of dark. Since we’re merging two layers, rather than using the dodge/burn brush, the extremes won’t become more extreme. If it still doesn’t look even enough you can always try adding an additional cloned layer. I recommend not merging the layers together until the very end, and always clone from the original image.


The difference is hardly noticeable. It’s not perfect, but it is better and far easier than trying to capture the same moment again.

I’ve been making some rookie mistakes lately with my photography. Remember when I went down to campus for some DIY graduation photos and forgot to change my aperture? I confess it hasn’t been the only time. At home, I can always re-take a photo, so I don’t sweat it. It’s a learning process. Since I want to enjoy my first vacation in two years, and not make it all about photography, I decided to shoot in aperture priority mode. That way, at least the exposure would be “right”. That was my intention. I went from indoors to outside with little thought. The shutter speed was no match for the large aperture and high ISO and ended up with some really overexposed photos. Woopsies.

Luckily I noticed fairly early in the process and also ended up with some great vacation photos, but this left me a question of what to do with the over exposures. I wanted to see if I could do anything with them. I immediately thought of my previous attempt at editing a photo to give it the appearance of a drawing. The first step was to dramatically increase the brightness like an over exposure. I decided to give it another go, this time following this tutorial I found. The tutorial was for photo shop, but I was able to replicate the steps pretty easily in paint shop pro. (I rarely see anyone using paint shop anymore, so I didn’t bother including the steps in this post. If you have any questions feel free to email me.)

The final product

The original photo (with an increase in contrast)

It was fast and simple, and I was happy with the results so I decided to try again with a few more photos. The correctly exposed photos turn out better than over exposed ones. (That’s usually the way it goes.) But with a few adjustments to contrast and brightness the exposure is correct enough that I can get something decent out of the photo. In some cases the glowing edges effect was too strong and I reduced the opacity, but overall the process did not involve much tuning. I was even able to finish this post during nap time.


The initial photo I tried to turn into a drawing.

While the initial tutorial was for creating a black and white drawing effect, I found I could produce decent colored drawings as well. I tried two different approaches, (1) one where I skipped the de-saturation step and (2) one where I left it in and reduced the opacity de-saturation layer to let some of the original color show through.

Method (1) Without the De-Saturation step

Method (2) With the de-saturation step, but the opacity of the de-saturation layer reduced

When skipping the de-saturation step in approach (1), the dodge layer effect is a little too strong in my opinion. The tutorial is designed for a black and white drawing that where the pencil strokes show the texture and edges of a shape. Solid blocks of color in the photo (like skin) appear mostly white. This makes sense in black and white, but less so in color. To correct this I reduced the dodge layer opacity to 96%. Now the ‘white’ regions have a touch of color to them.

I prefer the result from approach (2), where the opacity of the de-saturation layer is reduced. The overall color feels more even, and looks more like I would have expected from colored pencils. For the curious, the opacity of the de-saturation layer in the above photo is 66%, however I find this number to be highly variable based on the original image.

July 10, 2013

Photo Forensics

A while back I wrote about editing Nicki’s photos to reduce the appearance of orange in her nose nose. I hesitated before sharing that post, knowing I’d be opening myself up to criticism should anyone ever read my blog. What kind of mom would care about the color of her baby’s nose?, that’s who. I ultimately decided to publish the piece because I thought there might be others interested in doing something similar, and I wanted to be honest about the level of photo editing I do. My goal in editing these photos is to learn about photo editing, not (really) to change the image. Trying to keep the photo editing a secret feels like I’m sending the wrong message, like there’s something to be embarrassed or guilty about. Let me be clear here: I don’t think that there anything wrong with photo editing, nor do I think there is anything wrong with Nicki’s appearance. This is just the next step in learning to process photos, after color balancing and contrast/brightness adjustments.

The question I keep coming back to, however, is what level of photo editing is too much? In my mind if the photo edits are noticeable, it’s too much. If you can’t tell whether a photo has been altered, does it really matter if it has? (I have the same philosophy on real vs fake diamonds, for the record.)

Recently I’ve stumbled onto the field of photo forensics. I found and decided to have a go with my images. I use the full sized images (I only post small versions on my blog) since resizing them can affect the analysis.

De-oranging the nose: De-oranging is totally a word, by the way.

Screen Grab from FotoForensics.

No white patch, or higher ELA values around the nose! Looks like I can get away with this edit.

I also ran the analysis on the small (400 x 262) version I uploaded to my blog. Interestingly the nose appears unedited but the eyes show possible signs of being altered, where the above screen grab indicates they weren’t. I have brightened eyes before (using this tutorial), but not this photo. I double checked by going back to the raw file.

Removal of scratches:

At 7 months Nicki gave herself one heck of a scratch on her cheek at day care. (We referred to it as her daycare warrior battle scar). It was pretty pronounced for a while, and is still somewhat visible, although now it looks more like a blemish than a scratch. I’ve edited it out of a few photos where it was particular pronounced, like I would with baby acne and teething rashes. Although I’ve never been very good with teething rashes, they cover too much surface area.

The arrow shows where the scratch used to be pre edit. In this one I also brightened the eyes, but left the orange nose.


The birthmark:

Like the orange nose, I thought this one might be controversial. Nicki had a small birthmark on her left arm at birth. If it was a permanent birth mark, I would have let it be. Since it looked like something that would fade, and had the appearance of a bruise in photos, I removed it from some of her early photos. It was already fading by 2 months, but you can still see it in this instagram.

The arrow shows where the birth mark used to be pre edit. In this photo I didn’t completely remove the birthmark, just lessened it’s appearance.

Maybe? There’s a little bit of a lighter patch (higher ELA values) near where the birthmarked was removed, but it doesn’t stand out from other patches of skin that were unedited.

Other Photos

This is making me wonder if any of my edits would be detected. (Or, at least, detectable by me. I read the tutorials, but I’m no forensics expert.)


This photo was actually a composite of two photos. I took them early Christmas morning, when she and I were the only ones up. I can’t get Nicki to smile when I’m staring through the view finder, so I was singing and playing her while attempting to hold the camera level. The camera slipped slightly and the photo with the good smile was poorly framed. The nicely framed photo had lens flare on her face – over her left eye of all places! Eyes are the hardest to edit. What to do? Clone the face from the one photo onto the other.

It was fun running my photos through FotoForensics. Hopefully my analysis is not embarrassingly off. I want to learn more about forensics. And photo editing. And photo taking.

March 25, 2013

Carrot Nose

At Nicki’s last dr’s appointment her pediatrician commented that Nicki must “liked the orange vegetables”. Domingo thought that meant she was still wearing some of breakfast. (Confession Time: It wouldn’t have been the first time we had purees up the nose.) It wasn’t until I was taking photos of Nicki a few days later that I realized what the doctor was saying: Nicki is turning orange!


The recommendation when starting solid food is typically to stick with the orange veggies first, since they tend to be the most universally liked. The green veggies come next, followed by fruits, meats & yogurts. Gerber’s state 2 veggie options are: Carrots, Sweet Potatoes, Squash, Sweet Potatoes and Corn, Mixed Vegetable, Garden Vegetables, Green beans, and Peas. Mixed Vegetable? First ingredient is carrots. Garden Vegetables Mix? Second ingredient is carrots (and the third and final ingredient, Spinach, has beta carotene as well). That only leaves 2 non-orange options out of 9. Thinking about a meat puree? Unless it’s chicken apple, the first ingredient besides water is either sweet potatoes or… wait for it… carrots! There’s even beta carotene in breastmilk!

In other words, unless your child only eats fruits (and yogurt), there’s a good chance baby is getting a fair amount of beta carotene.

Nicki gets one vegetable, one fruit, one yogurt and rice cereal daily. We mix up the vegetables so she gets a pretty even mix. The orange veggies have lots of good nutrients necessary for her health, including beta-carotene, so I don’t really want to skimp on them in favor of more peas and green beans. Changing her diet for cosmetic reasons is beyond silly. The amount of beta carotene in her diet will decrease when she gets table food, but she’s not quite ready yet (my guess is next month) so we’re going to have to live with the orange hue a little while longer.

It’s not really noticeable. That is, until I get out the camera. The kind of lighting that brings out her beautiful dark eyes apparently also brings out that orange nose. With Easter and spring rapidly approaching, I know will be taking some more photos. I won’t have much time to do a lot of detailed work editing them (not that I know how to anyway), so I was on the hunt for a quick fix.

In the above photo the color of her nose is R:173 G:118 B:93. Her check is R:153 G:121 B:106. The Red/Blue ratio in her nose is nearly 2:1 where it’s 3:2 in her check. Green is also similarly way off. I need to add more blue and green to tone down the red. I created a second layer for the image, set my brush to R:192 G:255 B:255 and colored her nose. (I picked this color because it’s one of the 48 basic colors on my pallet. No real rhyme or reason there, just going for quick and simple.)


Next I decreased the opacity of the second layar down to 16%. Wallah! Just enough blue/green added so the red/blue and red/green ratios were not crazy off. It’s not perfect, but I think it’s pretty darn good quick fix.


As a side note, I’ve only done this to one other photo on my blog. Brownie points if you guess which one!

My new favorite photo of Nicki and I, after a fair amount of editing.

I fancy myself a semi-serious armature photographer. There are few things that make me happier than being behind my camera’s view finder. Heck, I took 3 thousand photos of Nicki before she was two months old. I have photos I adore of her. I have tons of her with Dada, Grandma, even the cat. But something is missing. Me! I have so few photos of Nicki and I!

This week I stumbled upon this on my phone.

Photo credit: my wonderful husband.
Nice composition, but the photo has the double whammy of slightly blurry (most noticeably by my hand), and nosy from high ISO (most noticeably on Nicki’s face.). I was probably rocking when the photo was taken.

Of course as soon as I saw it I wished I had a DSLR equivalent photo. I wanted something I could print out and frame. But alas, this was taken when Nicki was four weeks old. She’s a bit different today, and there’s no way we could recreate the photo. If I wanted a framed copy, I was going to have to fix it up in post processing.

For the first pass, I did some smoothing and contrast adjustment. While I can smooth out the texture a little bit, but there’s little I can do about the blur and loss of detail. My best bet is to make the issues look intentional, or at least as intentional as possible. Since the photo already has a warm, nostalgic feel to it, I decide to emphasize that and give in an old-time feel. To me that meant a Sepia filter (I picked Sepia over greyscale to preserve the warmth.) I used a special program to do this, but you can also do this by adjusting your hue and saturation to get the color palette you want.

Adjusting hue and decreasing saturation to simulate a sepia color contrast. In this case since the image is already pretty red, I didn’t have to adjust the hue much. Had there not been as much red in the photo, I might have had to adjust the three color channels (Red, Blue and Green) separately to get the right red/brown hues.

The next step was to clean up the background a bit. I did this by using the clone tool to get ride of the glass of chocolate milk, and adding a slight vignette to put the focus squarely on Nicki and I.

A comparison of the original (left) with the final version (right). Click it to see the comparison in more detail.

I’ve done a few photo enhancements in my time, but without a strong graphic design/drawing skill I’m limited to tools like the clone brush and minor adjustments. There seems to be an huge business opportunity here for someone with true graphic artists skills. How many of us have that one photo they would give there left arm to fix?