Nezibo Dreamception for iOS & Android harnesses Google Deep Dream Artificial Intelligence Neural Network

Adam Milton-Barker | Nov 3, 2015 | Artificial Intelligence | 3851

Back in July I wrote an article about Deep Dream, an Artificial Intelligence Convolutional Neural Network that Google uses to detect faces and patterns in photos uploaded to the platform. For those that haven't heard of Deep Dream yet, it is an A.I program that was developed for the I.L.S.V.R.C (ImageNet Large-Scale Visual Recognition Challenge), and was originally called Inception. The program was developed to detect patterns in images and classify the images based on its findings. During development it was found that the program actually creates psychedelic images which are its own interpretation of the images.

Not long after the article was published, Timothy Hands, the editorial content manager for a start up named Nezibo, in Farmingdale New Jersey, contacted me. At the time of contact there were a few ways that people could try out Deep Dream on the web, but waiting for your image to be processed was a painful experience. Of course you could download the code that Google had open sourced but this would require you being familiar with some complex programming protocols and experienced with the Python programming language.

Dreamception is the new app by Nezibo that is built using the A.I Neural Network. Nezibo had recently partnered up with, a US-based firm specializing in image rendering software & interfacing, and had also bought a cluster of AWS super computers and a GPU server to power their new creation. With the app Nezibo hoped to provide the power of Deep Dream in a fun to use application that can be downloaded on iOS and Android. The application allows you to take photos or choose from your gallery, then choose from a selection of predefined filters. Once you have chosen your filter you upload your image to the server and within seconds (depending on your connection) you will have an A.I's interpretation of your image.

The app is available free or paid. With the free version you get 10 filters and 10 edits each day, one filter on one image costs one credit and you can use multiple filters on one image. One cool feature is that if you share your images on social media you will actually receive credits back on to your account. The paid feature costs a very affordable $0.99 and allows you to have unlimited edits and also an additional 20 Deep Dream filters.

The app is not only for people to have fun with, Nezibo hope that the app will help spread awareness of Deep Dream technology, Timothy said:

"Deep Dream is exciting tech, and if people can learn about its potentiality moving forward and have fun with it at the same time, then mission accomplished!"

"We released Dreamception for two reasons: to spread the word about Deep Dream and make it accessible for everyone to play around with."

I totally agree, the technology behind Deep Dream is only the tip of the iceberg, in the following years Artificial Intelligence is going to be playing a major role in our every day lives, it already is behind the scenes, and Dreamception is a fun way of seeing some of the basic capabilities A.I can provide.

According to Nezibo, Dreamception iOS has been downloaded in over 22 countries and has over 50,000 downloads. The Android version which was released a little later on has had over double the downloads that the iOS version has had, one reason for this maybe support from the Android developer community.

I have had a play around on the iOS version and was very impressed, the app is easy to use and does what it says on the tin. You can check out my photos below, download the iOS Dreamception App here, the Android Dreamception App here or visit the official Dreamception website here.

I will be in contact with the guys at Nezibo regularly and look forward to seeing the progress of Dreamception and to seeing and testing out their future projects. You guys will be some of the first to know about developments from Nezibo in the future so keeps your eyes open for future info.