Lampix Whitepaper

Abstract

Project Overview
Quality Computer vision is an enabler for a very large set of applications, among which many are in the AR space. One such application is Lampix, a hardware and software solution that transforms any surface into a smart, augmented reality surface freeing human-computer interaction from digital screens. The Lampix platform is poised to be the foundation for the next generation of user experiences in retail, office, gaming, and many more environments by bridging the gap between the physical and the digital.

Computer vision and machine learning – Datasets
While the Lampix device is unique in its form factor, it relies on computer vision algorithms to recognize objects in the physical world and create augmented reality. This is similar to quite a few other AR systems (glasses, tablet-based, etc.), they all need to understand what they see before being able to augment that. This understanding requires large datasets of examples, such that algorithms (eg. Deep neural networks) can be trained to do the recognition.
While working on Lampix we realized the scarcity of such datasets. We were missing a general-purpose market for such datasets with a self-balanced demand & supply system. We always ended up gathering the data ourselves (and actually created software to do this comfortably on Lampix). But the idea of the general purpose market got stuck in our heads, and hence we started building the PIX ecosystem.

Where do we need help?
We are building the “PIX ecosystem”, which is the first blockchain-based ‘image mining’ network for augmented reality and other computer vision systems. To enable Lampix and other AR devices to reach their full potential we need to build a billion-image-and-description database, which will help in training algorithms for classification, detection and segmentation.

Why is this important?
Computer vision and machine learning are completely dependent on the availability of data, which is curated and described. We are therefore building the largest database of computer vision and machine-learning data composed of images with descriptions and segmentation maps.

Autonomous and fully automated
To build a large enough database that is meaningful it has to be built in an autonomous fashion. No central party has the means and capability of checking every single image and description for details. Therefore, voting and verification mechanisms that are distributed are the keys.

Decentralized
Data needs can be different from one environment to another. However, some data might actually be useful in more than one environment. The database relies on local providers and local consumers of the data. Our ecosystem can support both scenarios and will balance itself towards the most useful data. This would be most effectively and economically achieved through decentralization.

A market
For the database to be useful in a timely manner, we are building an efficient token-based market, which will reward people who provide data to the database and will charge people modestly for the usage of the data. We call this token as “PIX” which will be a digital currency used for the transactions within the market.
Lampix device is the first commercial use case for the market for computer vision datasets. While Lampix will be the easiest way to submit or consume data, any other computer vision or machine learning system will also be able to interact, through an open API, with the database and earn tokens or pay for the usage in tokens.

Existing traction and potential
With Lampix, our initial focus is on the enterprise sector, and we plan to do the same with the PIX ecosystem. Lampix is already generating revenue and great traction in this sector. Our first product on Lampix is an in-store product recommendation system being used by a large retail company. We are working on the first office applications together with Bloomberg and NYC Media Lab.

Lampix was accelerated at Highway1 in San Francisco (http://highway1.io/startup/lampix/) and the winner of AR & VR startup competition at SXSW (Festival, 2017). We were also featured at TechCrunch Disrupt and by a few other media outlets. Through this, we already have a pipeline of 185 companies many of whom are in the Fortune 100.
We will leverage this pipeline and extend it to the PIX ecosystem.

The token launch
We are doing a crowdsale of our token, which we call “PIX” at the end of July 2017. This will probably be the first “image mining, augmented reality” project built on top of Ethereum blockchain.

For image miners, we are going to charge PIX tokens to get a Lampix device and pay tokens to mine useful {image, description, segmentation} sets. We will also pay users in tokens to vote on these images. We use computer vision algorithms to approve or reject images before they are submitted to the vote. People will also use the tokens to purchase Lampix apps or earn PIX token by creating Lampix apps.

Our Vision
“Our vision is a Lampix in every home, office, store, and factory.”

Our Mission
“Our mission is to free humans from digital screens.”

Objectives
• To raise at least $60 million through a token crowd sale, for the development of the PIX ecosystem and Lampix as the reference platform for PIX
• To build a billion-image-and-description database by the end of 2018

Lampix Website
Lampix Whitepaper

Social