Computer Vision
Computer vision is a field of artificial intelligence that enables computers to interpret and process visual information such as images and videos in order to extract data. This data can then in turn be used to automate different processes. For example, In finance, computer vision can be used to automate document verification, fraud detection, and customer identification.
Computer vision that specialises in analysing textual information, such as a loan application or an ID, is often referred to as Optical Character Recognition or OCR.
Machine vision, visual data processing, image analysis
-
-
CV
Examples
A fintech startup offers a mobile banking app where new users can open an account in minutes. Instead of filling out lengthy forms on paper or online, users just take a photo of their government-issued ID as well as take a short selfie video. The app then uses computer vision to:
- Read and extract data from the ID (like their name, date of birth, and document number)
- Verify the ID’s authenticity by checking security features (e.g. holograms, MRZ lines)
- Match the selfie video against the ID photo to ensure the person is who they claim to be
This entire process happens in under a minute, replacing manual checks and reducing onboarding time by over 70%.
FAQ
How does computer vision work?
Computer vision systems process visual data through a series of steps: image acquisition, preprocessing, feature extraction, and interpretation. These systems utilise algorithms and learning models, often based on machine learning, to identify patterns and make decisions based on visual input.
What is computer vision used for?
Computer vision is employed across various industries to enable machines to interpret and process visual information. Common applications include facial recognition, self-driving vehicles, medical imaging diagnostics, quality inspection in manufacturing, and augmented reality experiences.
Is computer vision technology secure for handling sensitive financial data?
Yes, when used within a secure infrastructure. While computer vision models don’t encrypt or delete data themselves, they can be deployed on systems that encrypt data in transit and at rest, restrict access, and automatically delete or anonymise inputs after processing. Security depends on how the system is architected around the model, not just the model itself.
