Prisma Style Transfers

According to their website, the Prisma App has been installed over 100 million times. This App converts your photos into artwork using styles of famous artists like Van Gogh, Picasso, Levitan, as well as world famous ornaments and patterns. 

They use neural networks and artificial intelligence for creating timeless art from your own photos. I must admit that's pretty cool. But how do they do it? How do they create and release these image effects? And how do they make them work with any image?

My first hint was when they said they use artificial intelligence. Apple loves Prisma. They named Prisma the iPhone 2016 App of the year. And now Apple has released CoreML technology for working for both image recognition as well as image style transfers. CoreML uses this same kind of trained models for image recognition and image to art.  They call this machine learning. 

So where do these models come from? They have to be trained. You can just buy Prisma, but if you are into coding, you need to get your hands on transfer style models. There are technologies like Tensor Flow that can help. There's been a lot of progress in this area lately thanks to graphics processing units (GPUs). This has led to Deep Learning and the complex multi-level neural networks for creating models.

But there are two kinds of models. Those for image recognition and those for image style transfers. It is the style transfers that are used by Apps like Prisma. In order to create style transfer model, an artist or developer needs a style image and a content image. An example of a style image is the famous painting by Vincent van Gogh shown below. A content image is any image you want to start with. Once the model is created, users can transfer the style of any other image, not just the content image.

If you are working with IOS, you will need to convert the model to CoreML. There are various tools for doing this. The bottom line is that creative artists are able to use tools and existing art to create transfer models that are able to change the style of images into a form of art. The new art is similar to the style image used to create the model. What fun! 

I will be reviewing other Apps using AI models for image recognition and style transfer in future reviews. In the mean time, grab Prisma and have some fun. 

The Artificial Intelligence Myth

Check out this article  

It makes sense. Are we overusing the name Artificial Intelligence for marketing?

kauai Adventure

I'm back with new blog ideas. I was in Kauai doing research if you can believe that.

If it works, it works

The funny thing about programming is that if it works, it works and the debate is over. It doesn't matter how smart they are when they tell you it won't work. If it works, it works. But then they will ask if it's scaleable. 

Best Advice for Programmers Ever. The Dangerous Truth!

I love this article. It is the best advice for skilled programmers I have read in along time. The article makes a great point. It's worth reading

The fantastic truth is that code has become a Sanskrit like language

A Brave New World for Programmers

No nonsense Free Audio Playback of any Kindle or IBook

"I recently purchased a Kindle book and noticed that Audible offered a free audio version. But eventually I discovered it was not free and they had me signed up for a subscription. You know, those sneaky little leaks that gradually drain your bank account when you are not watching. Buyer Beware!"

I hijacked this article from here: Free Audio Playback of any Kindle or iBook

Here's a quick guide to enabling free, no subscription playback with page turning.

Open Settings and General->Accessibility->Speech->Speak Screen  then Enable Speak Screen. Below on the same screen, you are able to select from voices.  Then use a two finger swipe down from the top within your Kindle or iBook App.

IBM Watson Loves Apple to the CoreML

ET vs AI

We would love to review this App on Best Camera. 

Here is a good article for anyone interested in Artificial Intelligence, Aliens from other worlds, and fast Image Recognition. Will they provide the code for this App? I read in the article that they are planning to make the entire project open source on GitHub. That would be a great learning tool. And if they submit to the App Store, a good way to get a bunch of installs.

Here is a link for Part 2 of the Article. It looks like there are three parts.

Part 2 - Coding with Watson SDK and CoreML

There's kind of TOC in each article. I would recommend starting with Part 1, or just jump right to he coding with part 2.

Part 1 — Setting up Image Recognition Models in Watson Studio

What's great about this article is the in depth description of how to use some cool technology created with by an IBM/Apple partnership. This is a good read.

Quoting from the article:
The engineers at IBM and Apple have formed an exciting AI joint venture. This article will explain how these two giants have made your life, as a developer, easier and more productive.

The combo of Watson and CoreML has everything from friendly online model training tools to fast onboard image classification. It’s got the stuff we want.