Automated Graffiti Detection: A Novel Approach to Maintaining Historical Architecture in Communities
Bibliographic record
Abstract
Graffiti is common in many communities and even affects our historical and heritage structures. This leads to a decrease in the revenue associated with commercial activities or services (e.g., shops, restaurants, residences), and potentially reduces tourism in a region. Visual data, in the form of photographs, is becoming an efficient mechanism to record information. Photographs can be quickly captured, and are already frequently posted online by ordinary citizens (e.g., tourists, residents, visitors). Exploiting image data through automation and computer vision provides a new opportunity to simplify the current manual graffiti-monitoring processes, enabling automated detection, localization, and quantification of such markings. In this study, we developed a vision-based graffiti-detection technique using a convolutional neural network. Images collected from historical structures of interest within a community can be utilized to automatically inspect for graffiti markings. In the case in which citizens collect and contribute data, there is a high degree of duplication and repetition, and potentially a lack of GPS information. These hinder the direct use of the images for automating the process. To address these challenges, we built high-resolution, single-view façade images (orthophotos) before applying our robust graffiti detector. The robust graffiti detector was built using a database with 1022 images of damaged or contaminated structures gathered during a recent European Union project, entitled “Safeguarding Cultural Heritage through Technical and Organisational Resources Management” (STORM). A total of 818 images were used for training (10% of the training set was randomly chosen for the validation set), achieving 88% accuracy among the remaining 204 samples for testing. Using the trained detector, the technique developed was demonstrated using data collected from the Church of Agios Nikolaos (Leontariou), Kantza, Greece.
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 | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 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".