Segmentation and Clustering

From images to objects

Extracting objects

Image Segmentation

Intelligent Scissors (demo)

Intelligent Scissors [Mortensen 95]

Intelligent Scissors

Path Search (basic idea)

How does this really work?

Defining the costs

Defining the costs

Defining the costs

Dijkstra’s shortest path algorithm

Dijkstra’s shortest path algorithm

Dijkstra’s shortest path algorithm

Dijkstra’s shortest path algorithm

Dijkstra’s shortest path algorithm

Segmentation by min (s-t) cut [Boykov 2001]

Grabcut    [Rother et al., SIGGRAPH 2004]

Is user-input required?

Automatic graph cut [Shi & Malik]

Segmentation by Graph Cuts

Cuts in a graph

But min cut is not always the best cut...

Cuts in a graph

Interpretation as a Dynamical System

Interpretation as a Dynamical System

Color Image Segmentation

Extension to Soft Segmentation

Histogram-based segmentation

Histogram-based segmentation

Clustering

Break it down into subproblems

K-means clustering

K-Means++

Probabilistic clustering

Mixture of Gaussians

Expectation maximization (EM)

EM details

EM demo

Applications of EM

Problems with EM

Finding Modes in a Histogram

Mean Shift [Comaniciu & Meer]

Mean-Shift

Mean-shift for image segmentation

Choosing Exemplars (Medoids)

Taxonomy of Segmentation Methods

References