Graphixlang. A batch image processing DSL

Abstract

Nowadays, the demand for quality image processing for large volumes of content in domains like web development, e-commerce, photography, social media management and data science is constantly increasing. Image processing tasks like cropping, resizing, changing format tend to become repetitive and error prone if done by hand repeatedly. As the volume of content grows, so do the operational costs and inconsistent results, which eventually leads to lack of productivity. Previous research on batch image processing is focused on either complex performance optimized solutions or command-line scripting solutions, therefore resulting in a gap for users in the need of both simplicity and power. The purpose of this paper is to introduce a new batch image processing solution: GraphixLang, that prioritizes simplicity, readability, and is accessible to both technical and non-technical users, while still offering advanced capabilities for professionals. Therefore, the DSL ensures the automation of the imagery preprocessing pipeline for web development and other domains, with the aid of normalization, resizing, and augmentation of thousands of images in a batch.