AI-ordlista
Den kompletta ordlistan över AI
Instance Segmentation
Technique distinguishing each individual instance of objects from the same class in an image.
Panoptic Segmentation
Unified approach combining semantic and instance segmentation for comprehensive scene understanding.
Segmentation with FCNs
Fully Convolutional Network architecture enabling pixel-wise classification with feature maps.
Segmentation with U-Net
U-shaped architecture with skip connections preserving spatial details for medical segmentation.
DeepLab Segmentation
Family of models using dilated convolutions and atrous spatial pyramid pooling for precise segmentation.
3D/Volumetric Segmentation
Extension of semantic segmentation to three-dimensional data such as medical scans or point clouds.
Segmentation with Transformers
New generation of models using transformer architectures to capture global pixel relationships.
Active Learning Segmentation
Method optimizing data annotation by selecting the most informative samples to label.
Segmentation by Domain Adaptation
Technique allowing the adaptation of segmentation models between different domains without complete annotation.
Contour Segmentation
Approach focused on the precise detection of boundaries and limits between objects in the image.
Multi-scale Segmentation
Method that simultaneously processes multiple resolutions to capture both fine details and the global context.
Segmentation by Clustering
Unsupervised approach grouping similar pixels into regions without prior learning.