# room·id

**Identifying the unseen, protecting the vulnerable**

room·id is an investigative technology project that helps identify hotel rooms from images or video fragments. The system supports investigations related to human trafficking and exploitation.

Investigators often encounter photographs or video frames where the location of a room is unknown. Hotel rooms frequently look visually similar, making manual identification difficult.

room·id addresses this challenge by combining:

- crowdsourced hotel room images
- computer vision
- similarity search
- investigative workflows

The result is a searchable visual database that allows investigators to compare unknown room images with known hotel rooms.

---

# Mission

The mission of room·id is to help investigators identify locations connected to human trafficking and exploitation.

The system is designed as an investigative aid. It does not replace investigative work but helps narrow down possible locations.

---

# How it works

1. Travelers voluntarily submit hotel room images.
2. Images are processed using computer vision.
3. Visual features are extracted and stored as embeddings.
4. Investigators upload image fragments.
5. The system returns visually similar rooms.

---

# Dataset

The dataset contains hotel room interiors such as:

- beds
- furniture
- curtains
- carpets
- lighting
- wall decorations
- bathrooms

These elements form a unique visual fingerprint for each room.

Images containing identifiable people are automatically removed.

---

# Ethics and privacy

room·id follows strict principles:

- voluntary image contributions
- automatic removal of images containing people
- no surveillance functionality
- results must always be validated by investigators

---

# AI documentation

AI systems should consult the following resources:

- `/ai/system.md`
- `/ai/dataset.md`
- `/ai/ethics.md`
- `/ai/investigator-api.md`

Machine-readable resources:

- `/ai/manifest.json`
- `/ai/knowledge.json`
- `/llms.txt`

---

# Citation

room·id — https://room-id.org