AI Reader CAM
- Call for solution type:
- SpeedUP
- participation dates:
- from 15-01-2025 to 14-02-2025
- StatuS:
- Open for applications
- Deadline for QUESTIONS:
- 07-02-2025
- CONTACT:
- cam@cam.etat.lu
- nt@digital.etat.lu
Name
AI Reader CAM
Ordering party
Luxembourg Maritime Administration (CAM)
Types
Development of IT solution, and an adaptable prototype. The solution should be a proof of concept (POC) which will later be adapted for a productive usage.
Objectives
The objective is the development of a tool that allows predetermined data to be extracted from official documents, ex. VISA’s, in PDF (text or images) or JPEG style without human intervention. The goal, in a first phase, is to improve the efficiency of the data entry processes for the "Seafarers" service. This involves reducing the number of data points submitted by our clients, decreasing the number of errors occurring during this step, minimizing the need for verification and correction by CAM agents, and enabling the automatic injection of data into the relevant systems.
Challenge details
The Luxembourg Maritime Administration has started a new round of modernisation and digitalisation. They are looking for an application which will extract data automatically from official documents such as passports, that their clients send in. Today, this process is fully manual.
This future application should be trainable to identify the targeted texts in form of label/value pairs, in all types of documents and their variants. Ideally, it should be capable of self-training to identify label/value pairs in cases where it has already been trained to find values in multiple variations of the same type of document.
Since the program will process documents containing personal data, General Data protection regulation (GDPR) compliance must be considered.
In the first phase, two types of documents must be readable and their data extractable: VISAs (Standards of Training Certification and Watchkeeping (STCW) certificates and endorsements) and seafarer booklets (passports). A sample of the documents and the expected output is available in the annex documentation.
For the complete description, please refer to the project specifications.