10 - Image Segmentation

These flashcards cover key concepts related to image segmentation in computer vision, focusing on definitions and important methods discussed in the course.

Segmentation

A common task in image analysis and computer vision aimed at identifying meaningful regions within an image.

Region growing

A segmentation method that starts with a small ‘seed’ and expands the region by adding similar pixels until no more can be added.

Split and merge

A segmentation process where regions are split into smaller segments if they show too much variation and adjacent regions are merged if they are consistent.

Watershed algorithm

A technique used in edge-based segmentation that labels pixels based on their connectivity and the gradient of the image.

Gradient Image

Represents the rate of intensity change in an image, helping to locate edges and boundaries through watershed segmentation.

SLIC

Simple Linear Iterative Clustering, an effective method for generating superpixels that efficiently group similar pixels.

Clustering

A segmentation approach that seeks groups of similar pixels without regard for their spatial arrangement in the image.

Edge-based segmentation

A method that focuses on identifying edges in images, which represent discontinuities in image intensity.

Quadtrees

Data structures used to partition a two-dimensional space by recursively subdividing it into four quadrants or regions.

Intensity or Colour Characteristics

Local image properties used to partition or group pixels in image segmentation.