Posts Tagged as: photoshop

August 29, 2016

How neural networks create images

This is fascinating stuff.

According to Gene Kogan on Twitter: DGN “deep generator network” is a neural net that synthesizes realistic images by learning low-dimensional encoding over an image set and samples from it to activate neurons associated w/ a class, e.g. cheeseburger.

More examples here.

This is some seriously future shit.

August 17, 2015

How to Make a Fillhouette

I discovered something pretty cool in Photoshop. If you select the inverse of a person’s silhouette and apply a “Content-Aware Fill”, the results can be funny, interesting, and sometimes quite beautiful. And it literally takes about 15 seconds.

Here’s how to make a “fillhouette”.

1: Find an image that’s not too difficult to silhouette. Anything with well-defined edges of the person is good.

2: Select the person. The easiest way is to use the “Quick Selection Tool”.

3: Shrink the selection so it’s a few pixels inside the person’s silhouette. Go to “Select > Modify > Contract…” and choose a value around 3.

4: Now invert the selection so the background is selected instead of the person. Go to “Select > Inverse”.

5: Now, just go to “Edit > Fill…” and choose “Content-Aware” from the dropdown. Press OK and you’re all set!

Here are some other fillhouettes I just quickly put together.

I’d love to see what the creative community can do using this technique, so I started a Tumblr where I’ll be posting the best examples I come across:

February 5, 2014

Upstairs Downstairs

Here’s the view from the elevator in my new building – from the basement and the first floor. Mesmerizing and nauseating at the same time!


February 3, 2014

The image exceeds the size Save for Web & Devices was designed for.

Granted, this Photoshop dialogue box was probably conceived in the mid-to-late 2000s when mobile devices were barely in their infancies, but it’s particularly ironic seeing it today when I’m trying to export a photo I took with my iPhone.

Screen shot 2014-02-03 at 3.42.18 PM

June 26, 2013

Rasterize Layer

Wish I’d thought of this.

Via Reddit

June 6, 2013

120 New Yorker Cartoons Combined

What happens when you write a Photoshop action that combines hundreds of cartoons from The New Yorker? This!

May 6, 2013

Confusing User Interface Elements

Too many people are trying to make the web more usable so I thought I’d try the opposite. Below are some user interface elements I designed to confuse, badger, and infuriate users. Enjoy.

1: The Chutton
Part checkbox, part button. Ideally this will initially present itself in the “checked” state, regardless of user intent.

2: The Boggle
Functions as a regular checkbox, but activating the slider simply covers up the checkbox and label. Avoid action-oriented labels such as OK or YES, and instead use vague language and/or the past tense.

3: The Notifuckation
Displays the number of notifications alongside an active radio button. Clicking or tapping on the radio button immediately marks all the notifications as “read”.

4: The Search & Destroy
A standard text entry field, but the user must slide the search button to the left to submit the form. As the search button slides, it deletes the characters it touches.

5: The Homer
This UI element has everything; pre-populated tags that are confusingly named, 2 scroll buttons, a search field with a settings icon, and Windows controls that do nothing!

I’ve also made the PSD available for download. Please use irresponsibly.

April 11, 2013

Computers and Hollywood

I found this great subreddit about those designs on computer screens you see in movies. Some really interesting stuff in there.

Speaking of fake UIs, here’s a cool video made by designer Phillip Lentz. I’d love to open it fullscreen on my work computer and start frantically typing and see how long it takes someone to notice. If they ask what the hell I’m doing I’ll probably say something about how I’m just breaking through the encryption shield before uploading a nano-virus.

April 5, 2013

Listening to a Record from a Photograph

This blew my mind. Apparently it’s possible to use a photograph of a record and get an audible recording from it.

The step-by-step process illustrated in this article called Extracting Audio from Pictures makes sense once it’s broken down. Basically you convert the round image of the grooves into a straight line…

…then you use Photoshop to boost the contrast of the lines, which become the actual waveforms you feed into ImageToSound software.

Amazing stuff. I often wonder about future generations being able to eavesdrop on people in the past by using leaps in technology.

For example on this Wikipedia page about Archaeoacoustics:

Gregory Benford’s 1979 short story “Time Shards” concerns a researcher who recovers thousand-year-old sound from a piece of pottery thrown on a wheel and inscribed with a fine wire as it spun. The sound is then analyzed to reveal conversations between the potter and his assistant in Middle English.

The idea that people’s voices would have unknowingly been recorded onto the bowl like a modern day record, locked away until a future generation figures out how to extract the audio, is awesome. True, Mythbusters busted this myth, but I wouldn’t fully discount the general concept of learning more from old data with better and better tools. (Think the dino DNA in Jurassic Park.)

Taking that pottery example further, I wonder if when I take a photo with my iPhone, a future generation will be able to determine my mood that day by examining whatever minuscule image artifacts my brainwaves caused in the image sensor.

It sounds crazy, but not much crazier than one ancient potter telling his buddy to keep his voice down so the future can’t eavesdrop on their clay.

February 3, 2013


This helps guide me daily.