Photozilla: A Large-Scale Photography Dataset and Visual Embedding for 20 Photography Styles

TL;DR

This paper introduces 'Photozilla', a large-scale photography dataset with 990,000 images across 10 styles, and develops models for style classification and adaptation to new styles with minimal samples.

Contribution

The paper presents a new extensive dataset for photographic styles and a Siamese network approach for classifying and adapting to emerging styles with few samples.

Findings

Achieved ~96% accuracy in style classification.

Developed a Siamese network that identifies new styles with over 68% accuracy using only 25 samples.

Demonstrated the dataset's utility for style recognition and adaptation.

Abstract

The advent of social media platforms has been a catalyst for the development of digital photography that engendered a boom in vision applications. With this motivation, we introduce a large-scale dataset termed 'Photozilla', which includes over 990k images belonging to 10 different photographic styles. The dataset is then used to train 3 classification models to automatically classify the images into the relevant style which resulted in an accuracy of ~96%. With the rapid evolution of digital photography, we have seen new types of photography styles emerging at an exponential rate. On that account, we present a novel Siamese-based network that uses the trained classification models as the base architecture to adapt and classify unseen styles with only 25 training samples. We report an accuracy of over 68% for identifying 10 other distinct types of photography styles. This dataset can be found at https://trisha025.github.io/Photozilla/