Please note: The algorithm descriptions in English have been automatically translated. Errors may have been introduced in this process. For the original descriptions, go to the Dutch version of the Algorithm Register.
CATCH (Central Automatic TeChnology for Recognition of Persons)
- Publication category
- High-Risk AI-system
- Impact assessment
- DPIA
- Status
- In use
General information
Theme
Begin date
Contact information
Link to publication website
Responsible use
Goal and impact
The algorithm is designed to automatically compare faces. This is important for the police because it helps find out who people are in investigations. CATCH plays a big role in this. Recognition and identification of people are the basis for many police investigations. If a face in a photograph is unknown or there are doubts about who it is, investigations are carried out to find out who that person might be. By comparing the face with images from the Criminal Justice Database (SKDB), CATCH can help identify the person.
Police may compare faces if it helps in tracing crimes. In the case of an unknown suspect, a search in CATCH can be conducted. CATCH makes it easier to compare faces carefully and helps to quickly find a clue about the person's identity. This speeds up police work, increases the chances of finding the suspect and makes society safer.
Considerations
It is too much work for a human to compare large numbers of faces manually. Therefore, comparing faces is now done automatically. This increases the chances of recognition and makes the process more careful and faster.
Human intervention
After the algorithm automatically compares faces, CATCH creates a list of possible matches. An expert looks at this list and compares the faces. If the expert finds no matches, a second, independent expert looks at the list. This second expert does not use the findings of the first expert.
If the first expert finds similarities, this is seen as a possible recognition. This is called a possible match and is investigated further. The investigation is done by two experts working independently. The expert who first made the comparison does not participate in this further investigation. Both experts carry out a detailed analysis and record their findings. They look at specific features of the faces, such as the shape of the nose, lips and ears. Then they give their own judgement on the similarities or differences.
The final judgement is the joint judgement of both experts, based on their analyses.
Risk management
- The results of the algorithm are always checked by police experts according to a set working process.
- To avoid bias, the experts do not know which case they are investigating.
- The experts cannot see each other's results to ensure fairness.
- The experts do not know from each other who is working on the same case, so everyone remains impartial.
- If there is a possible match, an investigation is done by two other experts, independently of each other.
- The result of the facial comparison can only lead to a clue for further investigation.
- Only the trained and authorised experts at the Centre for Biometrics have access to the algorithm and the faces.
Legal basis
The processing of the comparison of biometric data falls under Police Data Act (Wpg) section 13(1c): identification of persons or things. Article 5 Wpg (special categories of police data) is also relevant regarding processing.
Taking photographs of suspects is done on the basis of Article 55c paragraph 2 and 4c of the Code of Criminal Procedure (WSv):
- Suspect serious offence (Art 67(1) Sv crime): Art. 55c paragraph 2 Sv gives investigating officers the obligation to take fingerprints and photographs.
- Suspect minor offence (no 67 paragraph 1 Sv offence): Doubt about identity? Art 55c paragraph 3 Sv gives the (assistant) public prosecutor the obligation to give an order to the investigating officer to take fingerprints and photographs. No doubt about identity? No possibility of taking fingerprints and photographs.
Links to legal bases
Elaboration on impact assessments
For CATCH, the police have prepared a Data Impact Assessment (GEB).
Impact assessment
Operations
Data
CATCH uses the Multi Biometric Identification System (MBIS). This system contains a database of biometric data of people (such as faces or fingerprints) and traces (such as DNA).
The database consists of three types of data:
1. Case data: These data determine what the biometric data may be used for, such as in which investigation. They also indicate how long the data may be kept.
2. Biometric data: These are biometric data of suspects, convicted persons, people who were suspected but not prosecuted, unknown suspects (such as traces) and deceased crime victims.
3. Biographical data: These are administrative data linked to the biometric data. They indicate who the person is, such as name and address.
Technical design
In CATCH, images of faces stored in a special database are converted into a code (biometric data). The algorithm compares the code of the face of an unknown suspect, witness or victim with the codes of faces stored in the Criminal Justice Database (SKDB).
External provider
Similar algorithm descriptions
- Tool that assists front desk staff in establishing identity by comparing the face of the person reporting to the photo on an identity document. Helps prevent look-alike fraud.Last change on 3rd of November 2025, at 8:50 (CET) | Publication Standard 1.0
- Publication category
- Other algorithms
- Impact assessment
- Field not filled in.
- Status
- In use
- Tool that assists front desk staff in establishing identity by comparing the face of the person reporting to the photo on an identity document. Helps prevent look-alike fraud.Last change on 3rd of November 2025, at 9:37 (CET) | Publication Standard 1.0
- Publication category
- Other algorithms
- Impact assessment
- Field not filled in.
- Status
- In use
- Tool that assists front desk staff in establishing identity by comparing the face of the person reporting to the photo on an identity document. Helps prevent look-alike fraud.Last change on 18th of September 2025, at 12:57 (CET) | Publication Standard 1.0
- Publication category
- High-Risk AI-system
- Impact assessment
- Field not filled in.
- Status
- In use
- Tool that assists front desk staff in establishing identity by comparing the face of the person reporting to the photo on an identity document. Helps prevent look-alike fraud.Last change on 26th of June 2025, at 9:59 (CET) | Publication Standard 1.0
- Publication category
- Other algorithms
- Impact assessment
- Field not filled in.
- Status
- In use
- Tool that assists front desk staff in establishing identity by comparing the face of the person reporting to the photo on an identity document. Helps prevent look-alike fraud.Last change on 3rd of November 2025, at 9:40 (CET) | Publication Standard 1.0
- Publication category
- Other algorithms
- Impact assessment
- Field not filled in.
- Status
- In use