Driving License Extractor


Driving License Extractor


An AI solution which allows real-time data extraction from driving licenses all over UK.

Pain Points

In many industries - especially banking, insurance, finance, public administration, etc. - verifying customers' identity (also known as Know Your Customer verification or KYC verification) is time and cost-consuming when done in the traditionally manual process. The major impediments of manual KYC verification are:

  1. Data entry: Human-prone errors (i.e. incorrect data input, inconsistent data format input); Costs to hire data entry personnel; System costs; etc.
  2. Data verification: Cost to manage and store customer's identification documents; Cost to hire data verification personnel; Human-prone errors; etc.
  3. Data storage and management

Offer Description

Our Driving License Extractor allows you to capture, review, verify and analyze identification information in real-time, from UK driving licenses under various formats (.png, .jpg, .jpeg, .pdf).
With an average processing time of just under 40 seconds per input and up to 97% accuracy, our solution can continuously process thousands of driving licenses and help cut the operation costs for your company by up to 60%.


  • Reduce operation costs by up to 60%.
  • Minimize errors in data entry or data loss.
  • Speed up the verification process.
  • Capture data in real-time

Our Technology

Our Driving License Extractor is a combination of advanced image processing techniques and natural language processing (NLP):

  • Image pre-processing: handles driving licenses with poor quality.
  • OCR technology: captures information on driving licenses accurately.
  • NLP technology: extracts information and understand the captured data (name, license number, date of issue, etc.).

Execution Guideline

After creating an Azure Virtual Machine using the provided VM image, connect to the VM and use the Command Line Interface to execute the model:

  1. Step 1: After connecting to the VM, you will be in a user folder. To access the 'home' directory, please use these commands:
    `$ sudo su Nexus`
    `$ cd ..`
  2. Step 2: In the 'home' folder, you will see the folder 'Nexus' which contains our model. To access the model folder, please use this command:
    `$ cd Nexus/driving-license-extractor`
  3. Step 3: Upload your Driving License image(s) onto the designated 'input' folder (acceptable extension: .png, .jpg, .jpeg, .pdf)
  4. Step 4: Run the below command to execute the model:
    `$ python3 pipeline --input=input/ --output=output/`
  5. Step 5: Access the 'output' folder using the path specified in the above command. You should find a result file in JSON format.


  1. The current version does not have a load-balancing feature. Therefore, we do not recommend multiple users using the tool at the same time