AI glossary · Fundamentals
Computer vision
In one sentence
Computer vision is the branch of AI that lets a system interpret images and video: recognising objects, reading text, spotting defects or describing a scene. Today it mostly relies on neural networks trained on large numbers of images.
Also known as: machine vision · image recognition · visual AI · CV
What is computer vision?
To a computer, an image is just a grid of numbers: the colour of each pixel. Computer vision turns those numbers into useful information: “this is an invoice”, “the total is this”, “this part has a crack”.
Put it to work
UTIO's AI, through an API
From €29/mo + VAT
For developers and agencies building their own products.
See the AI APIFor years it was done with hand-written rules: detect edges, measure shapes, match against templates. That worked in tightly controlled settings. The leap came with deep learning: networks trained on millions of images that work out for themselves which features set one object apart from another, even in poor light or from a different angle.
Related to this term
AI API
- Transcription, voice, text, image, OCR and embeddings
- Per-client keys with a spending cap
- Prepaid wallet: €50 minimum top-up
From €29/mo + VAT Panel account, plus usage
See the AI APIHow it works
- Capture. A photo, a scan or a video frame comes in as a grid of pixels.
- Feature extraction. A neural network works through the image layer by layer: edges and textures first, then shapes, then whole objects.
- Task. Depending on the model, the output is a classification (what is it?), a detection (where is each thing?), a segmentation (which pixels belong to which object?) or text reading.
The newest models are multimodal: they combine vision and language, so you can show them a photo and ask about it in plain English. The reverse journey is image-generating generative AI, which creates a picture from a description.
Like any model, computer vision gets more wrong when images are very different from what it saw in training: blurry photos, crumpled documents or unusual layouts.
Example
A plumbing and heating firm in Swansea receives hundreds of supplier delivery notes a month, many of them photographed on a phone in the van. Until now, someone typed each one into the accounts software by hand.
With computer vision, each photo goes through an OCR system that recognises the text and understands the layout: supplier, date, part numbers, quantities and amount. The data arrives already organised. A person only checks the ones the system flags as uncertain, for example when the photo is blurred.
Computer vision at UTIO
UTIO’s AI API includes OCR: it turns scanned invoices, delivery notes and forms into structured text, using the same key you use for transcription, text to speech, text, image and embeddings. It also generates images from a description.
A common pattern is to chain the pieces together: OCR extracts the data from the document, a text model classifies or summarises it, and the result is saved in UTIO’s CRM, which is also available through the API. A scanned document becomes data your team or your agents can actually work with.
Frequently asked questions
What's the difference between computer vision and image processing?
Image processing changes the image: cropping it, brightening it or removing noise. Computer vision interprets it: it tells you what's in it. The two are often combined: first the image is cleaned up, then a model analyses it.
Is OCR a type of computer vision?
Yes, it's one of the oldest and most widespread applications: recognising the text in an image or scanned document. Modern systems don't just read the characters; they also understand the document's layout, such as tables, fields and totals.
What are everyday examples of computer vision?
Face unlock on your phone, number plate recognition at a car park barrier, searching your photo library by what's in the pictures, scanning documents with your phone camera and quality checks on a production line.
Related terms
- OCR
OCR (optical character recognition) is the technology that turns the text in an image, a scanned PDF or a photo into digital text you can edit, search and process. With AI, it also understands the document's layout and structure.
- Deep learning
Deep learning is a type of machine learning that uses neural networks with many layers. Each layer picks out more abstract features of the data, which lets it handle text, speech and images without hand-written rules.
- Multimodal AI
A multimodal model is an AI model that can take in or produce several types of data, such as text, images, audio or video, and connect them. For example, it can read a photo of an invoice and answer questions about it.
- Neural network
An artificial neural network is a mathematical model made of layers of connected units, called neurons, that turn a numeric input into an output. It learns by adjusting the weight of each connection based on examples.
- Generative AI
Generative AI is artificial intelligence that creates new content, such as text, images, audio, video or code, from an instruction, using models trained on huge numbers of examples of that kind of content.
Updated on