Smartclip is a contextually-aware paperclip processor that crops and scales your images to maximize aesthetic quality.

While initially written as a passion project for The Glass Files, I've abstracted away all the hardcoded specifics so it should work generically. Images smaller than the desired thumbnail size are automatically padded with a border rather than being enlarged.

The algorithm is fairly simple and is an adjusted a port of smartcrop.js by Jonas Wagner:

• Find edges
• Find regions with a color like skin
• Find regions high in saturation
• Generate a set of thumbnail candidates
• Rank candidates using an importance function to focus the detail in the center and avoid it in the edges.
• The highest ranking candidate is selected and is processed by Paperclip

The process is super simple, and is basically identical to using any other Paperclip processor:

  # Your path might be different, so adjust as necessary
  require "#{Rails.root}/lib/paperclip_processors/smartclip"
  
  ...

  styles: { thumb: { resize_width: 150, resize_height: 150, processors: [:smartclip] } }

• Wrap within gem for easy installation
• Optimize pixel iteration to reduce number of loops. Can it be done without screwing with saliency patterns?
• Port and integrate face detection algorithm
• Completely rewrite to favor deep neural networks? Benefits from more complicated criteria, and can be trained from professional data rather than guesses
• Extract optimzation logic, along with what I've learned, to create separate image optmization processor?

If you found a bug, submit an issue. If you have some optimization insight, submit a Pull Request.