Development of AI and Human Based Artwork Classification System for Vision Transformer
Bibliographic record
Abstract
Rapid progress of generative models like DALL•E, Midjourney, and Stable Diffusion has rendered AI-created artworks more than ever hard to tell apart from pieces created by humans. This new phenomenon has created new challenges for digital authenticity and media forensics, especially as a visual gap between synthetic and human-created art has narrowed. In this study, I examine if contemporary deep learning architectures can indeed reliably determine the source of a work of art on a large variety of styles and visual intricacies. Using Kausthub Kannan’s AI vs Human Art dataset, three models Vision Transformer (ViT-B/16), ResNet50, and EfficientNet-B0 were trained at uniform conditions to carry out binary classification. To my knowledge, this work reports one of the earliest comparative methods of transformer-based vs convolution-based systems for AI-art detection in both an in-distribution and cross-generator scenarios. The experimental results show that ViT-B/16 performance is best achieving 98.49% performance and superiority over CNN baselines. Moreover, ViT-B/16 yielded greater toughness when tried on unseen DALL•E-generated images and out-of-distribution human artworks, indicating its patch-based global representation provides relevant benefits for this goal. These results illustrate the increasing importance of transformer architectures for digital art authentication and highlight the importance of evaluating detection models far beyond a single generative source, especially as AI-produced images come to be virtually indistinguishable from human artistic production.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".