Glossary
Computer vision
Systems that interpret images and video, inspection, counting, safety monitoring, reading.
Applied computer vision is mostly a data and conditions problem. Lighting, camera placement and labelling quality decide accuracy far more than architecture choice.
Precision and recall trade against each other, and where you set that balance is a business decision about the cost of a false reject versus a missed defect.
Data collection deserves its own plan. Images gathered casually, different cameras, inconsistent angles, varying light, produce a model that works in the conditions you happened to capture and fails in the ones you did not. A short, disciplined collection protocol at the start is worth more than any later modelling effort.
Commonly misunderstood: A single headline accuracy number usually hides wide variation between defect classes. Per-class reporting is the honest form.
Related terms, in context
The concepts you almost always meet alongside computer vision.
- Edge inference
- Running models on local hardware near the data, instead of calling a cloud API.
- OCR
- Extracting text from images and scans, including Indian scripts and handwriting.
- Object detection
- Locating and classifying objects within an image, returning boxes rather than a single label.
Where this shows up in our work
Computer vision is not an abstraction for us. It is a decision we make on live projects. It shows up most directly in computer vision, video analytics, where getting it wrong has a cost someone can measure.
If you are evaluating a vendor on this, the useful question is not whether they can define the term. It is what they measure, what they would refuse to do, and what happens in their system when the assumption behind computer vision stops holding.
Questions
What is Computer vision?
Systems that interpret images and video, inspection, counting, safety monitoring, reading.
What do people get wrong about computer vision?
A single headline accuracy number usually hides wide variation between defect classes. Per-class reporting is the honest form.
Does Orqent Labs build this?
Yes, Computer Vision Systems and Video Analytics. We work across India, covering all 19,238 PIN codes remotely.
Building something that involves computer vision?
We will tell you honestly whether it is the right approach for your problem.
Or email bd@dtrasglobal.com · call +91 74118 77878