/image-segmentation

Segmentation of a set of images by performing a contour-based segmentation on a reference proposed by an expert.

Primary LanguagePython

Contour-based segmentation

Open in Streamlit

Initial Demo

part1.mp4

Description

We want to perform the segmentation of a set of images by performing a contour-based segmentation on a reference proposed by an expert. Description

Generally, image segmentation goes through 3 phases:

  • Pre-processing (filtering)
  • Processing (contour segmentation)
  • Post-processing (morphology)

For each phase, there is a set of possible operators (methods or algorithms). Our task is to find and render the best operators that give a result very close to the ground truth. The tests are sometimes done on the operators and other times on their parameters.

Two types of images are given: Simple Images (Signals) and Textured Images (Real Images).