Page 19 - CT_AI_Class-7

It helps machines detect objects, individuals, movements or scenes by examining visual data
collected through cameras, sensors and software.

Computer vision is the process of teaching computers how to see and interpret images and
videos, enabling them to understand visual information.

Working of Computer Vision

Computer vision works by capturing visual data through cameras or sensors, then processing that
data to identify objects, people or patterns. The system analyses images and videos, extracting
useful information to enable machines to interpret and understand visual content.

Image Pre-processing Feature Extraction
Acquisition (Cleaning, Correction) (Shape, Color, Texture)

Object Categorisation Object Identification Object Tracking
(Classify) (Detect & Locate) (Follow)

Let us understand how the computer vision works:
1. Image acquisition: The computer captures an image using a camera, sensor or scanner,
similar to how we take photos or scan documents.

2. Pre-processing: The captured image is pre-processed by cleaning and enhancing it, which
includes removing noise, adjusting brightness and improving image quality for better
analysis.

3. Feature extraction: The system extracts important features from the image, such as edges,
shapes, colours and textures, to help in understanding the visual content.

4. Object categorisation: The system categorises the detected features into general classes,
such as identifying whether the object belongs to categories like a car, person or animal.

5. Object identification: The system identifies the exact object by matching it with stored data
or patterns to determine what the object specifically is.
6. Object tracking: The system tracks the movement of the identified object across multiple
frames or images, such as following a moving car or a walking person in a video.

AI Domains and Applications 17