المفاهيم الأساسية
The authors introduce the first multi-attribute, multi-category dataset specifically tailored for the aesthetic assessment of paintings and drawings, covering 24 distinct artistic categories and 10 aesthetic attributes. This dataset aims to catalyze advancements in the field of computational aesthetics for artistic images.
الملخص
The authors present the Aesthetics of Paintings and Drawings Dataset (APDD), a pioneering multi-attribute, multi-category dataset for the aesthetic assessment of paintings and drawings. The dataset was constructed with the active participation of 28 professional artists worldwide, along with dozens of art students.
APDD is structured into 24 distinct artistic categories based on different painting types, artistic styles, and subject matter. The dataset also includes 10 aesthetic attributes, such as theme and logic, creativity, layout and composition, space and perspective, sense of order, light and shadow, color, detail and texture, overall, and mood. The authors selected specific sets of attributes tailored to each artistic category.
The authors collected 4,985 paintings and drawings from various professional art websites and institutions, maintaining a 3:1 ratio between works by professional artists and student assignments to ensure diversity and representativeness. The images were then annotated by 51 professional annotators, resulting in over 31,100 annotation records.
To assess the aesthetic attributes and total scores of the paintings, the authors propose the Art Assessment Network for Specific Painting Styles (AANSPS), a novel approach designed for the evaluation of aesthetic attributes in mixed-attribute art datasets. The authors compare the performance of AANSPS with other state-of-the-art methods on the APDD dataset, demonstrating its effectiveness in predicting both total aesthetic scores and aesthetic attribute scores.
الإحصائيات
The APDD dataset contains a total of 4,985 images.
The dataset includes over 31,100 annotation records from 51 professional annotators.
Each image in APDD has been evaluated by at least 6 individuals.
اقتباسات
"The construction of a multi-attribute, multi-category dataset in the field of painting aesthetics represents a pioneering new task."
"Through close collaboration with approximately 60 global professional artists and students with high academic qualifications, we have successfully established a clear system for considering the aesthetic components of art images."