Tracemyfile brings you innovation and flexibility
Our system is built in Sweden, from scratch using the most modern design and technique.
Our system is lightweight, modular and scalable. It is not dependent on any third-party services or any specific cloud provider.
The innovative database structure makes it possible to handle up to a billion images per database server.
Tracemyfile system is highly dynamic, consisting of many different services - each operating on its own system, or launched as a group of tiny virtual mechanisms. Every service is designed to gracefully handle any other failure within the system.
The system consists of 5 distinct parts, each supported by one or several services:
We run neural networks on graphics processing units which creates a unique fingerprint for each image. This fingerprint is used both in similarity searches and to find exact or close matches – the copies derived from the original.
In order to make images searchable, we use the latest techniques for billion scale nearest-neighbour-search to compress and cluster the fingerprints and make them searchable within milliseconds.
Tracemyfile system is extremely flexible in where, how and when it crawls. It is not locked to any specific crawling strategy or technique, which makes it easy to adapt to our client’s particular needs by adding appropriate metrics for analyzing.
To learn more about our system or discuss your project feel free to contact us sales@eright.com.