Image and video analysis. Image segmentation

Today we are publishing the eighth lecture from the course “Image and Video Analysis” given by Natalia Vasilyeva at the Computer Science Center in St. Petersburg, which was created on the joint initiative of Yandex Data Analysis School, JetBrains and CS-club.



In total, the program has nine lectures, of which have already been published:

1. Introduction to the course "Image and video analysis" ;
2. Basics of spatial and frequency image processing ;
3. Morphological image processing ;
4. Building the signs and comparing images: global signs ;
5. Construction of signs and comparison of images: local signs ;
6. Similarity search. Search for fuzzy duplicates ;
7. Image and video analysis. Classification of images and object recognition .

Under the cut you will find a plan of the new lecture and slides.


')
What is a recognized object.

What is segmentation:

• Where segmentation is used;
• Subtasks segmentation.

Possible criteria for "community":

• Taxonomy of segmentation methods;
• Criteria "community". Colour;
• Criteria "community". Texture;
• Criteria "community". Location relative to the contour;
• Criteria "community". Movement, motion (motion);
• Criteria "community". Depth (depth);
• Criteria "community". Global.

Mathematical models:

• Use clustering;
• Clustering

The k-means method. Main idea:

• The k-means method. Algorithm;
• K-means method: step 1;
• K-means method: step 2;
• K-means method: step 3;
• K-means method: step 4;
• K-means method: step 5;
• Segmentation method of k-medium.

Adding Spatial Information:

• k-means: advantages and disadvantages;
• Mean-shift for image segmentation;
• Mean shift algorithm;
• Mean shift clustering / segmentation;
• Mean shift;
• Mean shift clustering;
• Mean shift segmentation results;
• More results;
• Mean shift: advantages and disadvantages;
• Probabilistic clustering;
• Expectation maximization (EM).

Hierarchical clustering:

• Model for metric space;
• Simulation using graphs;
• Automatic graph cut;
• Segmentation by Graph Cuts;
• Min cut;
• It is not always the best cut;
• Normalized Cut.

Segmentation examples:

• Use of graphs;
• Using 2-D lattices;
• Mathematical models;
• Top-down segmentation methods;
• Deformable contours;
• Parameterization;
• Contour energy setting;
• Optimization;
• Berkeley Segmentation DataSet [BSDS].