ICO Spotlight: Interview with George Popescu from Lampix

This post is part of the ICO Spotlight, a series of posts with the teams behind the different projects covered by Coinschedule.


Hello George, please start by telling us: What is Lampix and what is the Token Pix?

We believe we can free human computer interactions from digital screens. Think Tom Cruise in Minority Report. Imagine your room in 5 years: you will be able to use any surface around you as if it was a computer. The ability to transform any surface into an interactive computer (augmented reality) is going to unleash applications we have not even conceived of. Think about how much smartphones changed how we live. This is going to be a similar impact. A few examples of many include a retail store being able to have the same information as if you were on Amazon while having the product in your hands. You will be able to remotely collaborate and write with a co-worker’s paper remotely as if you were in the same room. Imagine using any table surface at a restaurant to order or play games. When cooking, you can place ingredients on a table, which would then display recipes and instructions next to them. Take a look at our youtube channel to see some examples in action.

Lampix is Augmented Reality without glasses. Lampix lowers the effort to develop AR apps for retail and other markets by offering a complete framework covering from computer vision to content.

Lampix transforms any surface into a smart surface. We make augmented reality without glasses by focusing on fixed horizontal surfaces. This makes AR much simpler. Lampix empowers any HTML5 page to react to objects, fingers, and movement, by means of a few simple attributes and a JavaScript library. Our training app then streamlines the learning of new object classes. This enabled all developers to make AR apps within hours. We are also working on an Android SDK with specific Lampix extensions to enable even more powerful apps.

 

Why do we need a database?
Any application that uses computer vision and machine learning (example: self-driving cars, Google Glasses, Hololens, MetaGlasses, other augmented reality systems, etc.) needs to be trained. Three types of data are needed: training, validation and testing. It is like training a human baby to talking: you need to show them, say what it is, make the baby repeat and verify that what they said is accurate or correct it. There are two parts in making the perfect computer vision system: a lot of data and good algorithms. It takes a lot of data in different contexts to train a good computer vision system. We believe using crowdsourcing and blockchain is the best approach to build such a database.

 

Why Crowdsourcing and blockchain?
Existing databases are controlled by the company who built them. Google for example has made such a database available here . However, there are two problems with this approach: Google controls this database and can at any moment forbid their competitors from using it, remove access to it, etc. Secondary, this database includes only the data Google thinks is needed.

We want a world where no single company, and not Lampix neither, will control who has access to the data, and what data should go in the database. No centralized control. This is why we are working on the blockchain. This database will be available to any miners or users to who earn or pay PIX tokens.

 

Why do we need tokens?
We are building a crowd-mined on the blockchain database of at least one billion sets. We believe this database should be independent of any central control. We believe the right mechanism to encourage mining is to pay miners for their labor. We believe we should also charge users a smaller amount for using the data in the database. PIX tokens can be spent on Lampix devices which will be the simplest way for miners to mine set. For all these reasons we need a token, we refer to as PIX, to help tying together this economy.

 

Wont there be issues running such a large database on top of a blockchain?

As things stand right now we do not expect to store all the data in the chain as it is not computationally effective. Our plan is to store hashes only of the data on the chain. However many new blockchains promise to deliver cheaper storage and computation so perhaps with new solutions this could become possible.
Our calculation using Amazon cloud storage show that 1 billion pictures at 4MB each would cost around $90,000 per month to store for example.

 

You have a goal of getting a billion pictures and descriptions. Why a billion, what happens when you reach this objective?

Based on experiences of companies like Google with their machine vision database 1-billion-data points seems to be a size that is meaningful and useful for a large set of systems and applications.

 

Tell us about Lampix (the product that originated the idea for the Token Pix). How is the product doing? Does it have paying customers?

Since we started Lampix in 2015 we have acquired paying clients like Bloomberg, PwC, large retail chains, large car manufacturers and other similar companies. Lampix was among the top 2% of companies who were accepted at the Highway 1 accelerator in Fall 2016. Lampix won best Augmented Reality and Virtual Reality company at SXSW festival, the largest in the world.

You can see here https://lampix.co/src/img/pipeline_company_list2.png a list of companies that are interested in working with Lampix.

 

What can you tell us about you and your team?

George is an MIT alumni based in New York and has 3 Masters degrees in Computer Science and Electrical Engineer, 3D Printing and Nanosciences. He was the founder of Boston Technologies (BT), the #1 fastest growing company in Boston in 2011. BT was also on the Inc 500/5000 list for 4 years in a row. He built 9+ companies to date. He lives in Hell’s Kitchen in New York. He is of Romanian origin.

Mihai is the Founder, CEO and CTO of Rosoftlab. Rosoftlab, since 2006, is a research and development firm in computer vision, augmented reality and related mobile and web apps. Their products and services are used by clients in Germany, Switzerland and many other countries. Employing more than 20 people, Rosoftlab counts among its clients Migros, the largest retailer in Switzerland. Rosoftlab’s continuous research and development has led to the creations of multiple companies, such as VanillaNAV, an easy to set up indoor navigation system.

 

Thank you so much George and good luck with your CTL/ICO!

Thank you!

Find out more about Lampix here