Graph Data Structure Overview

Images are pixels on a 2D grid. This is a type of densely connected graph, where each pixel in a quantized local region is a node and each neighbor is an edge. We can break images into regions using KMeans or SLIC. In this case we want to track neighboring relations between these regions. We convert images into a Graph datastructure consisting of Node nodes. Each Node is a collection of pixels representing a unique region. Each Node tracks its immediate neighbors. The Graph manages all Nodes.