Automated Cropping of Large Image Filesets with Python

Abstract

Autocropping of digital images offer an opportunity to both eliminate waste outside the region of interest and also to save on storage space for data. While work is ongoing on utilising machine learning techniques to autocrop images, the accuracy and repeatability of these techniques continues to pose challenges. This technical note describes a simple algorithm that can be used to autocrop large image filesets according to a pre-defined cropping criteria and is then is programmed using Python. The program is tested on two differing filesets and results are quantified in terms of storage space reduction. The results show that significant improvements in terms of the region of interest and in storage requirements can be attained with a simple approach.