Segmentation based object categorization

The image segmentation problem is concerned with partitioning an image into multiple regions according to some homogeneity criterion. This article is primarily concerned with graph theoretic approaches to image segmentation.

Applications of Image Segmentation

* Image Compression
** Segment the image into homogeneous components, and use the most suitable compression algorithm for each component to improve compression.
* Medical Diagnosis
** Automatic segmentation of MRI images for identification of cancerous regions.
* Mapping and Measurement
** Automatic analysis of remote sensing data from satellites to identify and measure regions of interest.

egmentation using Normalized Cuts

Graph theoretic formulation

The set of points in an arbitrary feature space can be represented as a weighted undirected complete graph G = (V, E), where the nodes of the graph are the points in the feature space. The weight $$w_{ij} of an edge $$(i, j) in E is a function of the similarity between the nodes $$i and $$j. In this context, we can formulate the image segmentation problem as a graph partitioning problem that asks for a partition $$V_1, cdots, V_k of the vertex set $$V, where, according to some measure, the vertices in any set $$V_i have high similarity, and the vertices in two different sets $$V_i, V_j have low similarity.

Normalized Cuts

Let G = (V, E) be a weighted graph. Let $$A and $$B be two subsets of vertices.

Let:

$$w(A, B) = sum limits_{i in A, j in B} w_{ij}

$$ncut(A, B) = frac{w(A, B)}{w(A, V)} + frac{w(A, B)}{w(B, V)}

$$nassoc(A, B) = frac{w(A, A)}{w(A, V)} + frac{w(B, B)}{w(B, V)}

In the normalized cuts approach [Jianbo Shi and Jitendra Malik (1997): "Normalized Cuts and Image Segmentation", IEEE Conference on Computer Vision and Pattern Recognition, pp 731-737 ] , for any cut $$(S, overline{S}) in $$G, $$ncut(S, overline{S}) measures the similarity between different parts, and $$nassoc(S, overline{S}) measures the total similarity of vertices in the same part.

Since $$ncut(S, overline{S}) = 2 - nassoc(S, overline{S}), a cut $$(S^{*}, {overline{S^{*}) that minimizes $$ncut(S, overline{S}) also maximizes $$nassoc(S, overline{S}).

Computing a cut $$(S^{*}, {overline{S^{*}) that minimizes $$ncut(S, overline{S}) is an NP-hard problem. However, we can find in polynomial time a cut $$(S, overline{S}) of small normalized weight $$ncut(S, overline{S}) using spectral techniques.

The Ncut Algorithm

Let D be an $$n imes n diagonal matrix with $$d on the diagonal, and let $$W be an $$n imes n symmetrical matrix with $$W_{ij} = w_{ij}.

After some algebraic manipulations, we get:

$$min limits_{(S, overline{S})} ncut(S, overline{S}) = min limits_y frac{y^T (D - W) y}{y^T D y}

subject to the constraints:
* $$y_i in {1, -b }, for some constant $$-b
* $$y^t D 1 = 0

Minimizing $$frac{y^T (D - W) y}{y^T D y} subject to the constraints above is NP-hard. To make the problem tractable, we relax the constraints on $$y, and allow it to take real values. The relaxed problem can be solved by solving the generalized eigenvalue problem $$(D - W)y = lambda D yfor the second smallest generalized eigenvalue.

The partitioning algorithm:
# Given a set of features, set up a weighted graph $$G = (V, E), compute the weight of each edge, and summarize the information in $$D and $$W.
# Solve $$(D - W)y = lambda D y for eigenvectors with the smallest eigenvalues.
# Use the eigenvector with the smallest eigenvalue to bipartition the graph.
# Decide if the current partition should be subdivided.
# Recursively partition the segmented parts, if necessary.

Example

Figures 1-7 exemplify the Ncut algorithm.

Limitations

Solving a standard eigenvalue problem for all eigenvectors (using the QR algorithm, for instance) takes $$O(n^3) time. This is impractical for image segmentation applications where $$n is the number of pixels in the image.

OBJ CUT

OBJ CUT [M. P. Kumar, P. H. S. Torr, and A. Zisserman. Obj cut. In Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, San Diego, pages 18-25, 2005.] is an efficient method that automatically segments an object. The OBJ CUT method is a generic method, and therefore it is applicable to any object category model.Given an image D containing an instance of a known object category, e.g. cows, the OBJ CUT algorithm computes a segmentation of the object, that is, it infers a set of labels m.

Let m be a set of binary labels, and let $$Theta be a shape parameter($$Theta is a shape prior on the labels from a Layered Pictorial Structure (LPS) model). We define an energy function $$E(m, Theta) as follows.

$$E(m, Theta) = sum phi_x(D|m_x) + phi_x(m_x|Theta) + sum Psi_{xy}(m_x, m_y) + phi(D|m_x, m_y) (1)

The term $$phi_x(D|m_x) + phi_x(m_x|Theta) is called a unary term, and the term $$Psi_{xy}(m_x, m_y) + phi(D|m_x, m_y) is called a pairwise term.An unary term consists of the likelihood $$phi_x(D|m_x) based on color, and the unary potential $$phi_x(m_x|Theta) based on the distance from $$Theta. A pairwise term consists of a prior $$Psi_{xy}(m_x, m_y) and a contrast term $$phi(D|m_x, m_y).

The best labeling $$m^{*} minimizes $$sum limits_i w_i E(m, Theta_i), where $$w_i is the weight of the parameter $$Theta_i.

$$m^{*} = arg min limits_m sum limits_i w_i E(m, Theta_i) (2)

The OBJ CUT algorithm

# Given an image D, an object category is chosen, e.g. cows or horses.
# The corresponding LPS model is matched to D to obtain the samples $$Theta_1, cdots, Theta_s
# The objective function given by equation (2) is determined by computing $$E(m, Theta_i) and using $$w_i = g(Theta_i|Z)
# The objective function is minimized using a single MINCUT operation to obtain the segmentation m.

Example

Figures 8-11 exemplify the OBJ CUT algorithm.

Other approaches

* Jigsaw approach [ E. Borenstein, S. Ullman: Class-specic, top-down segmentation. In Proceedings of the 7th European Conference on Computer Vision, Copenhagen, Denmark, pages 109-124, 2002.]
* Image parsing [Z. Tu, X. Chen, A. L. Yuille, S. C. Zhu: Image Parsing: Unifying Segmentation, Detection, and Recognition. Toward Category-Level Object Recognition 2006: 545-576]
* Interleaved segmentation [B. Leibe, A. Leonardis, B. Schiele: An Implicit Shape Model for Combined Object Categorization and Segmentation. Toward Category-Level Object Recognition 2006: 508-524]
* LOCUS [ J. Winn, N. Joijic. Locus: Learning object classes with unsupervised segmentation. In Proceedings of the IEEE International Conference on Computer Vision, Beijing, 2005.]
* LayoutCRF [J. M. Winn, J. Shotton: The Layout Consistent Random Field for Recognizing and Segmenting Partially Occluded Objects. CVPR (1) 2006: 37-44]

References