AI-DRIVEN AESTHETIC EVALUATION IN FINE ARTS: A MACHINE LEARNING APPROACH TO STYLE CLASSIFICATION

In the past, people with a lot of experience judged the beauty of fine arts by looking at them in the context of their deep cultural, political, and academic backgrounds. Now that artificial intelligence (AI) and machine learning (ML) are getting better, computers can better analyze and group artworks, making it possible to evaluate art in a way that is both scalable and objective. This study suggests a system for classifying styles in fine arts that is based on machine learning and combines both hand-made visual descriptions and deep learning-based feature extraction methods. The study uses a variety of datasets, such as WikiArt, Kaggle art collections, and selected museum records, to make sure that all types and movements of art are covered. To improve the quality of a dataset and lower its noise, preprocessing steps like colour normalization, cutting, and data addition are used. Feature extraction mixes common techniques like colour histograms, edge recognition, and texture analysis with deep features gathered from CNNs like VGGNet, ResNet, and EfficientNet that have already been trained. Transfer learning is used to make models fit the unique features of fine art images, which leads to better classification performance across a wide range of artistic fields. According to the results of experiments, hybrid feature fusion is much better at classifying things than single-method. It also gives us useful information about the visual elements that define different art styles. The suggested method can be used in systems for verifying, collecting, and suggesting art. It fills the gap between computer analysis and human-centered aesthetic judgement. This paper shows how AI could be used to help professional art critics do their jobs better, leading to progress in both computer vision and the fine arts.