Computer Vision

Computer vision is the field of AI that enables computers to interpret images and video, performing tasks such as classifying images, detecting and segmenting objects, reading text, estimating pose and depth, and tracking motion.

Classical computer vision used hand-designed features and geometry, such as edges, corners and stereo matching. Modern systems are dominated by deep learning, using convolutional neural networks and vision transformers trained on labelled images, and increasingly foundation models that support zero-shot recognition and segmentation. Core tasks include image classification, object detection, semantic and instance segmentation, optical character recognition, pose estimation, depth estimation and tracking.

Computer vision powers industrial inspection, robot guidance, worker safety monitoring such as detecting personal protective equipment or zone intrusion, traffic and logistics tracking, medical imaging, retail analytics and document processing. Machine vision is the industrial application of these techniques with engineered cameras and lighting.

Performance depends on representative training data, image quality and conditions such as lighting, occlusion and camera angle. Models are evaluated with metrics such as accuracy, mean average precision and intersection over union. Video analytics involving people raise privacy and data protection obligations, and edge deployment reduces latency and keeps images on site.

Key points

AI that interprets images and video
Core tasks are classification, detection, segmentation, OCR and tracking
Dominated by CNNs, vision transformers and vision foundation models
Video of people raises privacy and data protection obligations

Where AiVibe comes in

AiAmbA use cases include visual inspection on textile, automotive and electronics lines, robotics perception and worker safety.

Explore AiVibe’s work in AI & Machine Learning →

Computer Vision

Key points

Where AiVibe comes in

Related terms

Terms that refer to Computer Vision