Photo-Quality Evaluation based on Computational Aesthetics: Review of Feature Extraction Techniques

TL;DR

This paper reviews various feature extraction techniques used in computational aesthetics for photo quality evaluation, categorizing them by implementation complexity and discussing their application in real-world scenarios and future research directions.

Contribution

It provides a comprehensive taxonomy of feature extraction methods in aesthetic quality assessment, focusing on implementation complexity and practical integration.

Findings

Features are categorized by complexity and application context.

Machine learning results show promising accuracy in photo quality prediction.

Identifies unexplored areas for future research in computational aesthetics.

Abstract

Researchers try to model the aesthetic quality of photographs into low and high- level features, drawing inspiration from art theory, psychology and marketing. We attempt to describe every feature extraction measure employed in the above process. The contribution of this literature review is the taxonomy of each feature by its implementation complexity, considering real-world applications and integration in mobile apps and digital cameras. Also, we discuss the machine learning results along with some unexplored research areas as future work.