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.
Checking proportions of passport photo for a travel document
- Publication category
- Other algorithms
- Impact assessment
- Field not filled in.
- Status
- In use
General information
Theme
Begin date
Contact information
Link to publication website
Responsible use
Goal and impact
You come into contact with this algorithm when you apply for a passport or ID card. At the municipality, employees check whether the passport photo meets the requirements. The algorithm assists in checking four requirements.
- Are the eyes in the photo on a horizontal line?
- Is the head centred in the photo?
- Is the head not shown too small or too large?
- Is the head fully depicted?
The algorithm only performs measurements on proportions. Before this, that task was performed by employees themselves with a ruler.
If the photo does not comply even after human verification, the citizen has to submit a new photo.
Considerations
The quality of the photo in the travel document is important for checking a person's identity. This must therefore be done carefully. In doing so, an algorithm can support a Civil Affairs employee to improve and speed up that assessment. It helps them determine faster and more accurately whether the quality of the photo is good enough; does it meet the requirements? A compliant photo is in the citizen's interest. After all, it reduces the chances of errors in identification, for example at a border check. The alternative is for an employee to use a special ruler to check whether the photo meets the requirements. This takes more time and is less accurate.
This decision is not an automatic one. The clerk makes the decision after the result of the algorithm. The clerk may deviate from the algorithm in doing so.
Human intervention
This algorithm supports Civil Affairs employees in assessing the quality of a passport photo when applying for a travel document. This official ultimately makes the decision: does the photo meet the quality requirements? Even when the algorithm has rejected the photo, the officer can accept the photo for a number of reasons. The official can also still reject the photo after the algorithm approved it. For example, because of requirements that the algorithm does not check for. Think of a blurred photo or shiny spots on the face.
Reasons for accepting photos that do not meet the requirements:
- Physical or medical reasons;
- Someone is under six years old;
- Religious or philosophical reasons;
- Incorrect assessment of the algorithm.
These data are not recorded in this way for privacy reasons. Only recorded is: "Photo matrix exception" or "Application station exception".
The outcome of the algorithm is also not recorded, only displayed in the screen.
Risk management
A basic security level test was conducted on June 2022. No major risks to availability, security or integrity on services were found, using the photo algorithm.
Measures have been taken to avoid bias (advantages). The algorithm has been trained on different sources with photo material with different demographic characteristics. No specific problems are known either. A practitioner always has the option to approve a photo anyway if the algorithm advises otherwise.
Not recorded are the reasons for accepting photos anyway, after the algorithm has decided they do not meet the requirements. This is because of privacy risks surrounding that registration, as the reasons are, for example, medical or religious.
Legal basis
The Passport Act (section 3(2)) prescribes that a travel document must have a photo. For a good image of the applicant, the photo must meet a number of requirements. This is detailed in a number of Passport Implementation Regulations. For instance, the photo must be sharp, in colour, and the head must be fully visible. The photo matrix is based on the criteria set by the international standards of DOC9303 and ISO 19794:5.
This photo matrix can be downloaded here:
https://www.rvig.nl/fotomatrix
Exceptions have also been mentioned:
- You may cover your head for religious or philosophical reasons.
- You can deviate from the requirements for physical or medical reasons, sometimes this requires a statement from a doctor or medical institution.
- For children under the age of six, not all requirements need to be met.
That this photo matrix must be met is mentioned in the various passport implementation regulations:
- Passport Implementation Regulations Netherlands article 28
- Passport Implementation Regulations Caribbean Countries Article 40
- Royal Netherlands Marechaussee Passport Implementation Regulations Article 17
- Passport Implementation Regulations for Foreign Countries, Article 42
Operations
Data
Separate datasets are used for training, testing and validation. The training, testing and validation dataset contain representative data of the world population.
Technical design
The algorithm was created with support vector machine (SVM), a form of machine learning. After training and testing it, it was commissioned. The algorithm has since been fixed and does not continue learning on its own. The algorithm determines positioning partly based on the contrasts of pixels in the picture.
External provider
Similar algorithm descriptions
- The municipality conducts parking checks using scan cars. The cameras on the scan car use an image recognition algorithm to recognise license plates and check whether the vehicle is allowed to park there.Last change on 5th of January 2024, at 12:40 (CET) | Publication Standard 1.0
- Publication category
- Impactful algorithms
- Impact assessment
- Field not filled in.
- Status
- In use
- The municipality uses scan cars to perform parking checks. The cameras on the scan car use an image recognition algorithm to 1) recognize license plates and 2) check in the parking rights database whether the vehicle is allowed to park there.Last change on 12th of July 2024, at 9:26 (CET) | Publication Standard 1.0
- Publication category
- Impactful algorithms
- Impact assessment
- DPIA
- Status
- In use
- Automatic Number Plate Recognition cameras (ANPR) at four different locations record license plates of passing motor vehicles. The algorithm checks whether an exemption has been granted and determines whether photos are submitted for review in the enforcement system.Last change on 3rd of September 2024, at 6:04 (CET) | Publication Standard 1.0
- Publication category
- Impactful algorithms
- Impact assessment
- Field not filled in.
- Status
- In use
- Short description With the scan car, the municipality carries out parking checks. The scan car uses Automatic Number Plate Recognition (ANPR) cameras, identifies license plates and checks parking entitlements. The license plates without parking rights are transferred to the enforcement system for the purpose of collecting the parking tax due.Last change on 24th of November 2024, at 20:30 (CET) | Publication Standard 1.0
- Publication category
- Impactful algorithms
- Impact assessment
- DPIA
- Status
- In use
- The algorithm recognises the biometric data of driving licences, identity cards and passports, making it clear to the municipality whether the resident is who they say they are. This allows the municipality to check whether they are delivering municipal services to the right person.Last change on 26th of June 2024, at 19:05 (CET) | Publication Standard 1.0
- Publication category
- Impactful algorithms
- Impact assessment
- Field not filled in.
- Status
- In use