images.tooling.report

Automated psychovisual optimizations

Offers automated psychovisual optimizations (e.g. via machine learning) to increase compression without perceivable visual impact

a child looking into a telescope, pointed upwards

Introduction

All of these services and tools are ultimately about automating boring, repetitive work. With algorithms like SSIM, and libraries like Butteraugli being available for the better part of a decade, it has been possible to reduce the overwhelming amount of switches and toggles that image optimizing tools provide with the user-friendly alternative of “make this as small as possible, without changing how it looks”.

The Test

A feature examination was done to see if this was something offered. As this is, by its very nature, very computationally expensive, it was not expected to be performed by default. But rather as an option that a developer can enable should they desire.

Conclusion

akamai
bigCartel
cloudflare
cloudinary
ecwid
fastly
goDaddy
imageEngine

ImageEngine tracks the SSIM of the original and optimized image offline, and adjusts optimizations accordingly.

imaginary
imgix
imgproxy
optimole
shopify
shortPixel
smush
squarespace
thumbor
webflow
weebly
wix