https://github.com/LeviBorodenko/img2rag

Convert any image into a Region Adjacency Graph (RAG)
https://github.com/LeviBorodenko/img2rag

graph-algorithms image-processing image-segmentation rag region-adjacency-graph

Last synced: 3 months ago
JSON representation

Convert any image into a Region Adjacency Graph (RAG)

• Host: GitHub
• URL: https://github.com/LeviBorodenko/img2rag
• Owner: LeviBorodenko
• License: mit
• Created: 2020-04-15T18:02:32.000Z (about 4 years ago)
• Default Branch: master
• Last Pushed: 2020-04-27T15:37:45.000Z (about 4 years ago)
• Last Synced: 2024-03-27T18:05:00.267Z (4 months ago)
• Topics: graph-algorithms, image-processing, image-segmentation, rag, region-adjacency-graph
• Language: Python
• Homepage:
• Size: 248 KB
• Stars: 12
• Watchers: 3
• Forks: 1
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • Changelog: CHANGELOG.rst
  • License: LICENSE.txt

Lists

• awesome-stars - LeviBorodenko/img2rag - Convert any image into a Region Adjacency Graph (RAG) (Python)

README

        
## img2rag

Convert any image into its [Region Adjacency Graph](https://ieeexplore.ieee.org/document/841950) which can be used for either image segmentation or to create a graph embedding of the image.

![scheme](scheme.jpeg)



### Installation

Simply run `pip install img2rag`

### What is does

Given an image, we segement it into perceptually significant regions using first [Felzenszwalb segmentation](http://people.cs.uchicago.edu/~pff/papers/seg-ijcv.pdf) followed by a [threshold-cut](https://ieeexplore.ieee.org/document/841950). We then use the segmented regions to construct the following graph:

1. Each node corresponds to a segmented region.

2. We connect two regions if they are adjacent.

This is the so-called region adjacency graph. Furthermore, we add the following node-attributes to each region:

1. Location of the region centeriod

2. Orientation of the region

3. Mean and total color of the region

4. Size in px

The edges contain the mean-color difference between the two regions

# How to use

Simply import the `RAGimage` class and initiate with any image. Then use the build in methods to access various properties.

```python

from img2rag import RAGimage

# We assume the image is given as a numpy array or tf.Tensor with either 2 or 3 dimensions

# where the third dimension is the optional channel dimension.

img_tensor = [...]

# initiate RAGimage instance

image_rag = RAGimage(img_tensor)

# RAG as a networkx attributed DiGraph

image_rag.rag

# Scikit style labels of the image segementation

image_rag.labels

# Adjacency matric of the RAG

image_rag.adjacency

# Graph feature matrix of the RAG

# (Nodes x Node-Features)

image_rag.signal

```