Our algorithms are based on Machine Learning (ML). Our self-learnt neural networks have been exposed to over 10M images on the web, and have been trained to assign different bitrates to different parts of an image, therefore compressing them adaptively as opposed to a fixed block size (8×8 in JPEG for example).
We have replaced key components of the JPEG and PNG standards with deep learning based neural networks, and the resulting outputs are better in visual quality and compression rate compared to all existing methods.
In addition to compressing images to JPEG and PNG format, we also have a state-of-the-art ML implementation, creating our own filetype “*.mlvx” that guarantees compression below 1 bit-per-pixel whilst preserving the original quality and resolution.
We built our algorithms to be able to compress any spectral data, including LIDAR and Radio Frequency bands.
Speed up your websites and applications
Lowering the image file sizes means accelerated page load times in your websites and applications
Increase Search Engine Optimization (SEO) Rankings
Faster applications lead to higher rankings in search engine results
Faster load times combined with improved SEO increases conversions